Natural Instructions: Benchmarking Generalization to New Tasks from Natural Language Instructions

04/18/2021 ∙ by Swaroop Mishra, et al. ∙ 12

Can we enable NLP models to appropriately respond to instructional prompts and consequently generalize to new tasks? To study this question, we leverage the existing NLP datasets and the instructions that were used to crowdsource them to create NATURAL INSTRUCTIONS, a dataset of instructions and task-specific input/output data. This dataset consists of 61 distinct language instructions and about 600k task instances, and is used to evaluate existing state-of-the-art language-models (LMs) in addressing new tasks by few-shot prompting of GPT3 and fine-tuning BART. Our analysis indicates that: (a) the existing models indeed benefit from instructions and hence, show improved generalization to new tasks; (b) while models like GPT-3 generally benefit from instructions, the extent of their gains varies across different fields of instructions and also depends on the task being solved; (c) generalization to unseen tasks in NATURAL INSTRUCTIONS remains far from perfect for the state-of-the-art, indicating significant room for more progress in this direction.

READ FULL TEXT
POST COMMENT

Comments

There are no comments yet.

Authors

page 3

page 9

page 13

Code Repositories

natural-instructions-expansion

Expanding natural instructions


view repo
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.