Research
Datasets
All Projects
All Years
Viewing 1-10 of 73 datasets
Natural Instructions
A large benchmark of tasks and their language instructions
2022
The goal of Natural-Instructions project is to provide a good quality benchmark for measuring generalization to unseen tasks. This generalization hinges upon (and benefits from) understanding and reasoning with natural language instructions that plainly and…
Multihop Questions via Single-hop Question Composition
Multihop reading comprehension dataset with 2-4 hop questions.
Aristo • 2022
MuSiQue is a multihop reading comprehension dataset with 2-4 hop questions, built by composing seed questions from 5 existing single-hop datasets. The dataset is constructed with a bottom-up approach that systematically selects composable pairs of single-hop…
Drug Combinations Dataset
A Dataset for N-ary Relation Extraction of Drug Combinations
AI2 Israel • 2022
Combination therapies have become the standard of care for diseases such as cancer, tuberculosis, malaria and HIV. However, the combinatorial set of available multi-drug treatments creates a challenge in identifying effective combination therapies available…
NumGLUE
NumGLUE: A Suite of Fundamental yet Challenging Mathematical Reasoning Tasks
2022
Given the ubiquitous nature of numbers in text, reasoning with numbers to perform simple calculations is an important skill of AI systems. While many datasets and models have been developed to this end, state-of-the-art AI systems are brittle; failing to…
Web10K Dataset
38,176 queries and corresponding 1M+ images returned by Bing Image Search
PRIOR • 2022
Web10K is a dataset sourced from web image search data with over 10K concepts. It consists of 38,176 queries and the corresponding 1M+ images returned by Bing Image Search. Web10K provides dense coverage of feasible adjective-noun and verb-noun combinations…
The Fermi Challenge
A challenge dataset of Fermi (estimation) problems, currently beyond the capabilities of modern methods.
Aristo • 2021
A challenge dataset of Fermi (estimation) problems, currently beyond the capabilities of modern methods.
Qasper
Question Answering on Research Papers
AllenNLP, Semantic Scholar • 2021
A dataset containing 1585 papers with 5049 information-seeking questions asked by regular readers of NLP papers, and answered by a separate set of NLP practitioners.
BeliefBank
4998 facts and 12147 constraints to test a model's consistency
Aristo • 2021
Dataset of 4998 simple facts and 12147 constraints to test, and improve, a model's accuracy and consistency
EntailmentBank
2k multi-step entailment trees, explaining the answers to ARC science questions
Aristo • 2021
2k multi-step entailment trees, explaining the answers to ARC science questions
ATOMIC: An Atlas of Machine Commonsense for If-Then Reasoning
An atlas of everyday commonsense reasoning, organized through 877k textual descriptions of inferential knowledge.
Mosaic • 2021
We present ATOMIC, an atlas of everyday commonsense reasoning, organized through 877k textual descriptions of inferential knowledge. Compared to existing resources that center around taxonomic knowledge, ATOMIC focuses on inferential knowledge organized as…
1
2
•••
8
AI FOR THE COMMON GOOD
Email us: ai2-info@allenai.org
Call us: 206.548.5600
Follow us: @allen_ai
Subscribe to the AI2 Newsletter
© The Allen Institute for Artificial Intelligence - All Rights Reserved.
Privacy Policy
|Terms and Conditions

|Business Code of Conduct
AI2Allen Institute for AIAI2