data_source
stringclasses 1
value | prompt
listlengths 1
1
| response
stringclasses 1
value | reward_model
dict | ability
stringclasses 1
value | extra_info
dict |
|---|---|---|---|---|---|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Artificial Neural Networks (ANNs)**: Machine learning technique used to improve predictive analytics and decision-making by analyzing large volumes of data.\n\n## Relevant Research\n**EHR** - Key concepts: EHR\n**Attention Mechanism** - Key concepts: Attention Mechanism\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 264,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Webvid-2M**: A dataset of short videos with textual descriptions for training the vision-language branch.\n\n## Relevant Research\n**Semantic Cluster 41: bi-directional_lstm, lstm** - Key concepts: Semantic Cluster 41: bi-directional_lstm, lstm\n**Atomic Claims** - Key concepts: Atomic Claims\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 440,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Cosine Distance**: A metric used for measuring similarity between word embeddings.; A metric used to measure similarity in the CLIP embedding space.\n\n## Relevant Research\n**COCO** - Key concepts: COCO\n**Flash-Attention2** - Key concepts: Flash-Attention2\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 328,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Global Direction**: A method that maps text prompts into a single manipulation direction in StyleGAN's style space.\n\n## Relevant Research\n**Audiovisual Segmentation (AVS)** - Key concepts: Audiovisual Segmentation (AVS)\n**Retrieval** - Key concepts: Retrieval\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 486,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Distance to Goal (DTG)**: A metric measuring the distance between the agent and the target goal at the end of an episode.\n\n## Relevant Research\n**Data Augmentation** - Key concepts: Data Augmentation\n**Hyperparameters** - Key concepts: Hyperparameters\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 368,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**FID (Fréchet Inception Distance)**: Metric for assessing the quality of generated images.; A metric to measure the quality and realism of generated images in comparison to real images.\n\n## Relevant Research\n**L3MVN** - Key concepts: L3MVN\n**GRU (Gated Recurrent Unit)** - Key concepts: GRU (Gated Recurrent Unit)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 79,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Cube-Stacking Environment**: An environment where an RL agent stacks cubes in a target order.\n\n## Relevant Research\n**Mask2Former** - Key concepts: Mask2Former\n**Semantic Cluster 21: deep_reinforcement_learning, reinforcement_learning** - Key concepts: Semantic Cluster 21: deep_reinforcement_learning, reinforcement_learning\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 148,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Open-vocabulary semantic segmentation (OVS)**: A task aiming to segment images using arbitrary categories specified by labels or captions.\n\n## Relevant Research\n**GPT-3.5** - Key concepts: GPT-3.5\n**Question Generation** - Key concepts: Question Generation\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 530,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**R2R**: Room-to-Room dataset used for training and evaluating navigation agents.\n\n## Relevant Research\n**TrackGo: A Flexible and Efficient Method for Controllable Video Generation** - Key concepts: TrackGo: A Flexible and Efficient Method for Controllable Video Generation\n**Multi-Modality Arena** - Key concepts: Multi-Modality Arena\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 381,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Ontology**: A formal representation of knowledge as a set of concepts within a domain, and the relationships between those concepts, often used to enhance understanding in ZSL.\n\n## Relevant Research\n**MMA-Diffusion** - Key concepts: MMA-Diffusion\n**BARD** - Key concepts: BARD\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 319,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Multimodal Dataset**: nan\n\n## Relevant Research\n**SummEval** - Key concepts: SummEval\n**Content-insensitive model** - Key concepts: Content-insensitive model\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 776,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Dataset**: nan\n\n## Relevant Research\n**Truthfulness Measures** - Key concepts: Truthfulness Measures\n**Bert Family** - Key concepts: Bert Family\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 773,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Sim-to-Real Gap**: The challenge of transferring learning from simulation to real-world environments.\n\n## Relevant Research\n**RLHF** - Key concepts: RLHF\n**DBSCAN** - Key concepts: DBSCAN\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 133,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**LangNav**: A navigation system using language as a perceptual representation for vision-and-language navigation.\n\n## Relevant Research\n**Knowledge graph** - Key concepts: Knowledge graph\n**General OCR Theory: Towards OCR-2.0 via a Unified End-to-end Model** - Key concepts: General OCR Theory: Towards OCR-2.0 via a Unified End-to-end Model\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 380,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Seq2Seq**: A sequence-to-sequence model architecture used in natural language processing tasks, including semantic parsing.\n\n## Relevant Research\n**Technique** - Key concepts: Technique\n**Student Model** - Key concepts: Student Model\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 213,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Latent Space**: A space used in StyleGAN for image manipulation, including the W+ space.\n\n## Relevant Research\n**Semantic Cluster 63: attention_weights, attention_mechanism, visual_instruction_tuning** - Key concepts: Semantic Cluster 63: attention_weights, attention_mechanism, visual_instruction_tuning\n**QF-attack** - Key concepts: QF-attack\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 483,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Knowledge Graphs**: Structures representing medical knowledge and relationships, aiding in reasoning.; Structured knowledge bases that LLMs can utilize to improve factual accuracy.\n\n## Relevant Research\n**Self-supervised Learning** - Key concepts: Self-supervised Learning\n**Semantic Cluster 6: lora, lorahub, foundational_models...** - Key concepts: Semantic Cluster 6: lora, lorahub, foundational_models...\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 55,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Retrieval-Augmented Training Datasets (RATD)**: Datasets developed to help smaller reasoning models utilize relevant information from longer text sequences.\n\n## Relevant Research\n**Habitat Simulator** - Key concepts: Habitat Simulator\n**Cascaded Diffusion Models** - Key concepts: Cascaded Diffusion Models\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 426,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Task Completion**: A standard metric used to measure the success ratio of navigation tasks.\n\n## Relevant Research\n**Semantic Cluster 83: success_weighted_by_path_length, soft_success_rate** - Key concepts: Semantic Cluster 83: success_weighted_by_path_length, soft_success_rate\n**Hyperparameters** - Key concepts: Hyperparameters\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 606,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**CARLA**: A simulation environment used for autonomous driving experiments.\n\n## Relevant Research\n**Semantic Cluster 3: self-supervised_learning, zero-shot_learning, few-shot_learning** - Key concepts: Semantic Cluster 3: self-supervised_learning, zero-shot_learning, few-shot_learning\n**Behavior Cloning** - Key concepts: Behavior Cloning\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 601,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Open Information Extraction (OpenIE)**: A method for extracting open entities and relations from text data.\n\n## Relevant Research\n**Cost-Utility Approach** - Key concepts: Cost-Utility Approach\n**Velocity Control Commands** - Key concepts: Velocity Control Commands\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 554,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Textual Inversion**: An optimization-based image editing method that allows for rapid concept acquisition in text-to-image models without fine-tuning the entire model.\n\n## Relevant Research\n**HumanEval** - Key concepts: HumanEval\n**Matterport3D dataset** - Key concepts: Matterport3D dataset\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 620,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**BERT Model**: A deep learning model used for natural language processing tasks, specifically for classifying operations in queries.\n\n## Relevant Research\n**Semantic Cluster 36: examinee, examiner, factuality_evaluation...** - Key concepts: Semantic Cluster 36: examinee, examiner, factuality_evaluation...\n**Planning** - Key concepts: Planning\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 196,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Data Augmentation**: Techniques used to generate additional training data for improving model performance.\n\n## Relevant Research\n**LSTM** - Key concepts: LSTM\n**code-davinci-002** - Key concepts: code-davinci-002\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 294,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Soft Success Rate**: A metric weighted by path length to evaluate navigation performance beyond mere success.\n\n## Relevant Research\n**Content-sensitive model** - Key concepts: Content-sensitive model\n**COLMAP** - Key concepts: COLMAP\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 377,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**T5**: A transformer model for text generation and fine-tuning across various tasks.; A family of pre-trained language models utilized for summarization and evaluation tasks.; A text-to-text transformer model leveraged for automatic template generation and NLP tasks.; A pretrained Transformer-based language model used for finetuning, inference, and evaluation.\n\n## Relevant Research\n**Fine-grained Vision-Language Correlation** - Key concepts: Fine-grained Vision-Language Correlation\n**Computational Ontologies** - Key concepts: Computational Ontologies\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 60,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**BM25 index**: A retrieval system used to access oracle contexts in the NQ dataset.\n\n## Relevant Research\n**Clip Family** - Key concepts: Clip Family\n**BARD** - Key concepts: BARD\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 575,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Observation Encoding Model**: A model that encodes current visual observations into tokens for processing by LLMs.\n\n## Relevant Research\n**SQLNet** - Key concepts: SQLNet\n**Semantic Cluster 73: sam, maskcomp, mask2former** - Key concepts: Semantic Cluster 73: sam, maskcomp, mask2former\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 375,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**FLAN-T5**: A large language model used for experiments, known for its zero-shot and few-shot learning capabilities.\n\n## Relevant Research\n**Mobility Guidance Matrix** - Key concepts: Mobility Guidance Matrix\n**Sentiment Labels** - Key concepts: Sentiment Labels\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 286,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**GAIL**: Generative Adversarial Imitation Learning framework for training policies through expert demonstrations.\n\n## Relevant Research\n**Mutual Reasoning Makes Smaller LLMs Stronger Problem-Solvers** - Key concepts: Mutual Reasoning Makes Smaller LLMs Stronger Problem-Solvers\n**EXAMINEE** - Key concepts: EXAMINEE\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 602,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Simulation**: nan\n\n## Relevant Research\n**SAM** - Key concepts: SAM\n**Mask-Prompt Segmentation Datasets (DUTS-TE, ECSSD, GrabCut, Berkeley, DAVIS)** - Key concepts: Mask-Prompt Segmentation Datasets (DUTS-TE, ECSSD, GrabCut, Berkeley, DAVIS)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 788,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Input-First and Output-First Approaches**: Two strategies for generating task instances based on instructions.\n\n## Relevant Research\n**Instruction tuning** - Key concepts: Instruction tuning\n**ROUGE** - Key concepts: ROUGE\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 393,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Redpajama**: A dataset used for training models in this research.\n\n## Relevant Research\n**LoRA** - Key concepts: LoRA\n**Unseen Parallel Prosody Transfer (UPPT)** - Key concepts: Unseen Parallel Prosody Transfer (UPPT)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 247,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Perplexity**: A metric used to measure the uncertainty in predicting the next token in a sequence.\n\n## Relevant Research\n**LongLoRA** - Key concepts: LongLoRA\n**Vision Model** - Key concepts: Vision Model\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 585,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Shallow Fusion**: A technique used to combine ASR models with language models for improved decoding.\n\n## Relevant Research\n**WaveNet** - Key concepts: WaveNet\n**ControlNet** - Key concepts: ControlNet\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 296,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**DDIM (Denoising Diffusion Implicit Models)**: Models used for sampling and inversion in the inference phase.; Denoising Diffusion Implicit Model, a method employed in the denoising process of diffusion-based models.\n\n## Relevant Research\n**Semantic Cluster 92: video-chatgpt, vicuna, valley...** - Key concepts: Semantic Cluster 92: video-chatgpt, vicuna, valley...\n**History Collector Model** - Key concepts: History Collector Model\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 90,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Semantic Cluster 46: positional_interpolation, pose, rope...**: nan\n\n## Relevant Research\n**Building and better understanding vision-language models: insights and future directions** - Key concepts: Building and better understanding vision-language models: insights and future directions\n**Falsehoods Dataset** - Key concepts: Falsehoods Dataset\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 695,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**BloombergGPT**: A finance domain-specific language model designed for financial-related tasks.\n\n## Relevant Research\n**mPLUG-Owl3: Towards Long Image-Sequence Understanding in Multi-Modal Large Language Models** - Key concepts: mPLUG-Owl3: Towards Long Image-Sequence Understanding in Multi-Modal Large Language Models\n**Habitat** - Key concepts: Habitat\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 181,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Video Question Answering (VideoQA)**: Evaluates model understanding of videos based on questions.\n\n## Relevant Research\n**Video Object Segmentation and Tracking (VOST)** - Key concepts: Video Object Segmentation and Tracking (VOST)\n**LongLoRA** - Key concepts: LongLoRA\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 447,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**L3MVN**: A novel framework leveraging large language models (LLMs) for visual target navigation.\n\n## Relevant Research\n**Semantic Cluster 104: cross-entropy_loss, cross-entropy_loss_function** - Key concepts: Semantic Cluster 104: cross-entropy_loss, cross-entropy_loss_function\n**Audiovisual Segmentation (AVS)** - Key concepts: Audiovisual Segmentation (AVS)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 363,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Longformer**: A transformer model designed for processing long documents using a sparse attention mechanism.\n\n## Relevant Research\n**Atomic Claims** - Key concepts: Atomic Claims\n**Semantic Cluster 104: cross-entropy_loss, cross-entropy_loss_function** - Key concepts: Semantic Cluster 104: cross-entropy_loss, cross-entropy_loss_function\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 275,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Synthetic dataset**: A dataset created using the TrueTeacher method, consisting of\n\n## Relevant Research\n**Behavioral Cloning (BC)** - Key concepts: Behavioral Cloning (BC)\n**ScienceQA** - Key concepts: ScienceQA\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 158,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Curated Datasets**: Selected data sources used to train models for better accuracy in specialized tasks.\n\n## Relevant Research\n**DAVIS Dataset** - Key concepts: DAVIS Dataset\n**Referring Navigable Regions (RNR)** - Key concepts: Referring Navigable Regions (RNR)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 69,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Semantic Cluster 49: flash_attention, flash-attention2, attention_transition**: nan\n\n## Relevant Research\n**EAGLE: Exploring The Design Space for Multimodal LLMs with Mixture of Encoders** - Key concepts: EAGLE: Exploring The Design Space for Multimodal LLMs with Mixture of Encoders\n**Semantic Cluster 77: lseg, paintseg, relation-aware_intra-modal_matching** - Key concepts: Semantic Cluster 77: lseg, paintseg, relation-aware_intra-modal_matching\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 698,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Hindsight Relabeling**: A technique to improve agent learning by re-labeling past experiences.\n\n## Relevant Research\n**Speaker Identification** - Key concepts: Speaker Identification\n**COLMAP** - Key concepts: COLMAP\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 131,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Navigating Beyond Instructions: Vision-and-Language Navigation in Obstructed Environments**: nan\n\n## Relevant Research\n**Fréchet Inception Distance (FID)** - Key concepts: Fréchet Inception Distance (FID)\n**Semantic Cluster 79: habitat_simulator, matterport3d** - Key concepts: Semantic Cluster 79: habitat_simulator, matterport3d\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 44,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Hyperparameters**: Configuration settings governing the training of machine learning models.; Configuration settings influencing the behavior of learning algorithms.\n\n## Relevant Research\n**ToolBench** - Key concepts: ToolBench\n**Behavioral Cloning (BC)** - Key concepts: Behavioral Cloning (BC)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 70,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**RGB-D Sensors**: Sensors used for capturing depth and color information in the environment.\n\n## Relevant Research\n**Root Mean Square Error (RMSE)** - Key concepts: Root Mean Square Error (RMSE)\n**TrueTeacher** - Key concepts: TrueTeacher\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 613,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**13. LLaVA-Instruct-150K**: A previous vision-language instruction-following dataset, noted for its limitations in visual diversity and multi-modality.\n\n## Relevant Research\n**Selective Context** - Key concepts: Selective Context\n**LongCite: Enabling LLMs to Generate Fine-grained Citations in Long-context QA** - Key concepts: LongCite: Enabling LLMs to Generate Fine-grained Citations in Long-context QA\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 398,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Autoregressive Conditioning**: A method where the model predicts the next token based on previous tokens in a sequence.\n\n## Relevant Research\n**Spatial Super-Resolution (SSR)** - Key concepts: Spatial Super-Resolution (SSR)\n**Soft Success Rate** - Key concepts: Soft Success Rate\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 311,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Evidence Set Score**: A scoring mechanism in multi-hop retrieval systems to quantify the sufficiency of selected sentences for answering queries.\n\n## Relevant Research\n**Habitat-Matterport 3D (HM3D)** - Key concepts: Habitat-Matterport 3D (HM3D)\n**COIN** - Key concepts: COIN\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 431,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Behavioral Cloning (BC)**: Training models by imitating expert actions.\n\n## Relevant Research\n**Semantic Cluster 72: sample-weighted_focal_contrastiveloss, soft_hirschfeld-gebelein-rényiloss, focal_contrastive_loss** - Key concepts: Semantic Cluster 72: sample-weighted_focal_contrastiveloss, soft_hirschfeld-gebelein-rényiloss, focal_contrastive_loss\n**LongWriter: Unleashing 10,000+ Word Generation from Long Context LLMs** - Key concepts: LongWriter: Unleashing 10,000+ Word Generation from Long Context LLMs\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 135,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Spatio-Temporal Reward Feedback Learning (ST-ReFL)**: Algorithm that optimizes the video diffusion model using multiple reward models for improved video quality and motion consistency.\n\n## Relevant Research\n**TruthfulQA** - Key concepts: TruthfulQA\n**Semantic Cluster 75: zero-shot_classification, zero-shot_performance** - Key concepts: Semantic Cluster 75: zero-shot_classification, zero-shot_performance\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 82,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**SummEval**: A benchmark for evaluating faithfulness in summarization across multiple models and datasets.\n\n## Relevant Research\n**Evaluation** - Key concepts: Evaluation\n**Zero-shot approach** - Key concepts: Zero-shot approach\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 165,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**XSum**: A summarization dataset known for its more abstractive summaries that can lead to factual errors.; A dataset used for summarization tasks.\n\n## Relevant Research\n**Collision Avoidance** - Key concepts: Collision Avoidance\n**Semantic Cluster 95: latent_optimization, latent_mapper, textual_inversion** - Key concepts: Semantic Cluster 95: latent_optimization, latent_mapper, textual_inversion\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 164,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Better Alignment with Instruction Back-and-Forth Translation**: nan\n\n## Relevant Research\n**CCNet** - Key concepts: CCNet\n**NarrativeQA** - Key concepts: NarrativeQA\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 28,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Semantic Cluster 5: t5, t5_text_encoder, homegrid...**: nan\n\n## Relevant Research\n**Human-in-the-loop** - Key concepts: Human-in-the-loop\n**BERT** - Key concepts: BERT\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 654,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Question Generation**: The process of forming questions from specific claims to gather additional information.\n\n## Relevant Research\n**Zero-shot approach** - Key concepts: Zero-shot approach\n**SHOW-O: ONE SINGLE TRANSFORMER TO UNIFY MULTIMODAL UNDERSTANDING AND GENERATION** - Key concepts: SHOW-O: ONE SINGLE TRANSFORMER TO UNIFY MULTIMODAL UNDERSTANDING AND GENERATION\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 193,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**GovReport**: A dataset containing reports published by the U.S. Congress and Government, used for language modeling tasks.\n\n## Relevant Research\n**Semantic Cluster 2: knowledge_graphs, factscore, knowledge_graph...** - Key concepts: Semantic Cluster 2: knowledge_graphs, factscore, knowledge_graph...\n**Video-LLaMA** - Key concepts: Video-LLaMA\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 239,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Mask-Prompt Segmentation Datasets (DUTS-TE, ECSSD, GrabCut, Berkeley, DAVIS)**: Datasets used for evaluating mask-prompt and point-prompt segmentation.\n\n## Relevant Research\n**Lattes Platform** - Key concepts: Lattes Platform\n**LaGR-SEQ** - Key concepts: LaGR-SEQ\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 519,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Closed-loop Scientific Discovery**: An approach automating the entire scientific process, utilizing active learning, GPR, PCFG, and reinforcement learning.\n\n## Relevant Research\n**mFACE dataset** - Key concepts: mFACE dataset\n**LAW OF VISION REPRESENTATION IN MLLMS** - Key concepts: LAW OF VISION REPRESENTATION IN MLLMS\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 501,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Reinforcement Learning (RL)**: A learning paradigm where agents learn optimal behavior through rewards.; A learning paradigm where agents learn to make decisions through interactions with the environment.\n\n## Relevant Research\n**Semantic Cluster 68: graph_neural_network, graph_neural_networks** - Key concepts: Semantic Cluster 68: graph_neural_network, graph_neural_networks\n**Temporal Modeling Module** - Key concepts: Temporal Modeling Module\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 136,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Latent Mapper**: A network trained to infer manipulation steps in latent space for efficient image editing.\n\n## Relevant Research\n**MMLU** - Key concepts: MMLU\n**Nouveau-token Backdoor Attack** - Key concepts: Nouveau-token Backdoor Attack\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 485,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**CLIP Score**: A metric for evaluating content consistency in generated videos based on semantic coherence.\n\n## Relevant Research\n**API** - Key concepts: API\n**Semantic Cluster 93: ma-lmm, lvu, video_language_co-attention_network** - Key concepts: Semantic Cluster 93: ma-lmm, lvu, video_language_co-attention_network\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 536,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Vision-Language Navigation (VLN)**: A task where agents navigate environments based on natural language instructions.\n\n## Relevant Research\n**v-parameterization** - Key concepts: v-parameterization\n**Multi-Modality Arena** - Key concepts: Multi-Modality Arena\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 140,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**MPC-style controller**: A controller that replans after a set number of steps to improve motion smoothness.\n\n## Relevant Research\n**Collision Avoidance** - Key concepts: Collision Avoidance\n**Chain-of-thought prompting** - Key concepts: Chain-of-thought prompting\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 542,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Self-Taught Evaluators**: nan\n\n## Relevant Research\n**Human-in-the-loop** - Key concepts: Human-in-the-loop\n**Motion Prior** - Key concepts: Motion Prior\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 6,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**SUPER-NATURALINSTRUCTIONS**: A benchmark dataset used to evaluate the performance of instruction-tuned models.\n\n## Relevant Research\n**COARSE CORRESPONDENCES Elicit 3D Spacetime Understanding in Multimodal Language Model** - Key concepts: COARSE CORRESPONDENCES Elicit 3D Spacetime Understanding in Multimodal Language Model\n**Semantic Cluster 97: attribute_space, ideal_point** - Key concepts: Semantic Cluster 97: attribute_space, ideal_point\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 388,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Semantic Cluster 102: visual_prompt_tuning, bev**: nan\n\n## Relevant Research\n**TruthfulQA** - Key concepts: TruthfulQA\n**BLIP** - Key concepts: BLIP\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 751,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**T5 Text Encoder**: A large frozen language model used for conditioning inputs based on text prompts.\n\n## Relevant Research\n**GAP** - Key concepts: GAP\n**Root Mean Square Error (RMSE)** - Key concepts: Root Mean Square Error (RMSE)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 73,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Contextualized Calibration**: A technique to adjust predicted probabilities based on prior distributions of label words.\n\n## Relevant Research\n**Semantic Cluster 50: zeroscrolls, zero-shot_approach** - Key concepts: Semantic Cluster 50: zeroscrolls, zero-shot_approach\n**Seq2Seq** - Key concepts: Seq2Seq\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 558,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**MemeCap**: Dataset for generating and analyzing meme captions.\n\n## Relevant Research\n**CORECT** - Key concepts: CORECT\n**Framework** - Key concepts: Framework\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 457,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Secondary RL Agent**: An agent that determines when to query the LLM for solutions.\n\n## Relevant Research\n**Semantic Cluster 49: flash_attention, flash-attention2, attention_transition** - Key concepts: Semantic Cluster 49: flash_attention, flash-attention2, attention_transition\n**Feed-forward approach** - Key concepts: Feed-forward approach\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 145,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**SpecAugment**: A data augmentation method for speech recognition that operates on the log mel spectrogram, including techniques like Time Warping, Frequency Masking, and Time Masking.\n\n## Relevant Research\n**Longformer** - Key concepts: Longformer\n**Spatio-Temporal Reward Feedback Learning (ST-ReFL)** - Key concepts: Spatio-Temporal Reward Feedback Learning (ST-ReFL)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 291,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Contriever**: A widely used model for unsupervised dense information retrieval.\n\n## Relevant Research\n**External Knowledge** - Key concepts: External Knowledge\n**IEMOCAP** - Key concepts: IEMOCAP\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 234,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Deepeye**: An open-source software for organizing and ranking visualizations based on query results.\n\n## Relevant Research\n**Semantic Cluster 85: r2r, carla-nav** - Key concepts: Semantic Cluster 85: r2r, carla-nav\n**Component Matching** - Key concepts: Component Matching\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 220,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Attention Weights**: Used in the model to visualize and interpret predictions by highlighting the importance of different input frames.\n\n## Relevant Research\n**Messenger** - Key concepts: Messenger\n**Attention mechanism** - Key concepts: Attention mechanism\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 300,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Graph Neural Network (GNN)**: A neural network architecture designed to work with graph-structured data, useful for modeling inter-class relationships in ZSL.\n\n## Relevant Research\n**Imagen Video** - Key concepts: Imagen Video\n**DDIM (Denoising Diffusion Implicit Models)** - Key concepts: DDIM (Denoising Diffusion Implicit Models)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 316,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Vision-Language Models (VLM)**: Models integrating visual and language understanding for various tasks.; Models that learn rich vision-language correlations from web-scale image-text pairs, enabling zero-shot predictions across various visual tasks.\n\n## Relevant Research\n**Text-to-Text API** - Key concepts: Text-to-Text API\n**ClinicalGPT** - Key concepts: ClinicalGPT\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 132,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Fréchet Inception Distance (FID)**: A metric for evaluating image generation quality.\n\n## Relevant Research\n**PG19** - Key concepts: PG19\n**Direct-a-Video** - Key concepts: Direct-a-Video\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 526,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Domain Dataset**: nan\n\n## Relevant Research\n**LoraHub** - Key concepts: LoraHub\n**Video Object Segmentation and Tracking (VOST)** - Key concepts: Video Object Segmentation and Tracking (VOST)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 777,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Video-LLaMA**: A multi-modal language model for understanding visual and auditory content in videos.\n\n## Relevant Research\n**Operation Extraction** - Key concepts: Operation Extraction\n**RoPE (Rotary Position Embedding)** - Key concepts: RoPE (Rotary Position Embedding)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 436,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Semantic Cluster 96: prefgen, paired_comparison_queries**: nan\n\n## Relevant Research\n**Building and better understanding vision-language models: insights and future directions** - Key concepts: Building and better understanding vision-language models: insights and future directions\n**Language Task** - Key concepts: Language Task\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 745,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Student Model**: Smaller language models (e.g., ALBERT, DeBERTa, OPT-\n\n## Relevant Research\n**Prompt Augmentation** - Key concepts: Prompt Augmentation\n**Elastic Weight Consolidation (EWC)** - Key concepts: Elastic Weight Consolidation (EWC)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 425,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**To Code, or Not To Code? Exploring Impact of Code in Pre-training**: nan\n\n## Relevant Research\n**Multi-head Self-attention** - Key concepts: Multi-head Self-attention\n**Semantic Vectors** - Key concepts: Semantic Vectors\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 41,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**LINGO Algorithm**: A clustering algorithm used for labeling cluster instances and organizing texts into hierarchical thematic clusters.\n\n## Relevant Research\n**Policy Learning** - Key concepts: Policy Learning\n**Intelligent Document Processing (IDP)** - Key concepts: Intelligent Document Processing (IDP)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 494,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Velocity Control Commands**: Commands generated to control the movement speed and direction of UAVs.\n\n## Relevant Research\n**Object-Backdoor** - Key concepts: Object-Backdoor\n**TrueTeacher** - Key concepts: TrueTeacher\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 108,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Semantic Cluster 48: dragon, contriever**: nan\n\n## Relevant Research\n**GenAgent: Build Collaborative AI Systems with Automated Workflow Generation -- Case Studies on ComfyUI** - Key concepts: GenAgent: Build Collaborative AI Systems with Automated Workflow Generation -- Case Studies on ComfyUI\n**Language Metric** - Key concepts: Language Metric\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 697,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Self-supervised Learning**: A training technique allowing models to learn from large datasets without explicit labels.\n\n## Relevant Research\n**Video Q-former** - Key concepts: Video Q-former\n**StyleGAN** - Key concepts: StyleGAN\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 56,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**mBERT**: A multilingual version of BERT, a language representation model pre-trained on multiple languages.; A multilingual BERT model used for creating text-only embeddings.\n\n## Relevant Research\n**Semantic Cluster 108: json_format, string_format** - Key concepts: Semantic Cluster 108: json_format, string_format\n**RoPE (Rotary Position Embedding)** - Key concepts: RoPE (Rotary Position Embedding)\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 259,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Ai Model**: nan\n\n## Relevant Research\n**Elastic Weight Consolidation (EWC)** - Key concepts: Elastic Weight Consolidation (EWC)\n**Observation Encoding Model** - Key concepts: Observation Encoding Model\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 761,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**LAION**: A large-scale vision-text dataset that can contain noise and biases, affecting model training.\n\n## Relevant Research\n**RELIC** - Key concepts: RELIC\n**Knowledge transfer** - Key concepts: Knowledge transfer\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 409,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Retrieval-augmentation**: A method that combines retrieval techniques with language models to enhance performance on long context tasks.\n\n## Relevant Research\n**ELEVATER** - Key concepts: ELEVATER\n**LAW OF VISION REPRESENTATION IN MLLMS** - Key concepts: LAW OF VISION REPRESENTATION IN MLLMS\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 227,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**SQL**: Structured Query Language used for managing and querying relational databases.\n\n## Relevant Research\n**Simulation Environment** - Key concepts: Simulation Environment\n**Semantic Cluster 96: prefgen, paired_comparison_queries** - Key concepts: Semantic Cluster 96: prefgen, paired_comparison_queries\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 212,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Semantic Cluster 41: bi-directional_lstm, lstm**: nan\n\n## Relevant Research\n**InstructGPT** - Key concepts: InstructGPT\n**Hyperparameters** - Key concepts: Hyperparameters\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 690,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**LSTM**: Long Short-Term Memory, a type of recurrent neural network suitable for sequential data analysis.\n\n## Relevant Research\n**Cross-Entropy Loss Function** - Key concepts: Cross-Entropy Loss Function\n**Attention Transition** - Key concepts: Attention Transition\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 331,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**xGen-MM (BLIP-3): A Family of Open Large Multimodal Models**: nan\n\n## Relevant Research\n**Semantic Cluster 61: rnn_language_model, dialoguernn** - Key concepts: Semantic Cluster 61: rnn_language_model, dialoguernn\n**Global-to-Local Mechanism** - Key concepts: Global-to-Local Mechanism\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 24,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Adversarial Masked Contrastive Painting (AMCP)**: A process that alternates between inpainting and outpainting to refine segmentation masks.\n\n## Relevant Research\n**Language Metric** - Key concepts: Language Metric\n**AI and ML Techniques** - Key concepts: AI and ML Techniques\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 516,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Temporal-Distance IoU (TDIoU) Loss**: A loss function designed to enhance the training of temporal video grounding (TVG) models.\n\n## Relevant Research\n**Workflow Automation** - Key concepts: Workflow Automation\n**LAION-COCO** - Key concepts: LAION-COCO\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 350,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
|
scientific_discovery
|
[
{
"content": "# Scientific Discovery Task\n\nIdentify key limitations in the current research on these topics and propose innovative solutions to address them.\n\n## Key Entities\n**Fine-grained Vision-Language Correlation**: A focus on local correspondence knowledge to enhance performance in tasks like object detection and segmentation.\n\n## Relevant Research\n**Large Language Models (LLMs)** - Key concepts: Large Language Models (LLMs)\n**GLUE Benchmark** - Key concepts: GLUE Benchmark\n\n## Task Requirements\n1. Your proposal should combine at least two of the key entities in a novel way\n2. Explicitly identify limitations in existing approaches and how your proposal addresses them\n3. Include a clear methodology section with specific steps\n4. Consider cross-domain applications that span multiple research areas\n5. Structure your response with appropriate headings (Introduction, Background, Methods, Expected Results)\n\nWrite your scientific discovery proposal:\n",
"role": "user"
}
] |
{
"ground_truth": "",
"style": "rule"
}
|
Scientific_Discovery
|
{
"index": 355,
"split": "train",
"cross_domain": false,
"entities": [],
"papers": [],
"task_type": ""
}
|
End of preview. Expand
in Data Studio
README.md exists but content is empty.
- Downloads last month
- 3