This paper explores a simple method for improving the zero-shot learning abilities of language models. We show that instruction tuning -- finetuning language models on a collection of tasks described via instructions -- substantially boosts zero-shot performance on unseen tasks. 
We take a 137B parameter pretrained language model and instruction-tune it on over 60 NLP tasks verbalized via natural language instruction templates. We evaluate this instruction-tuned model, which we call FLAN, on unseen task types. FLAN substantially improves the performance of its unmodified counterpart and surpasses zero-shot 175B GPT-3 on 19 of 25 tasks that we evaluate. FLAN even outperforms few-shot GPT-3 by a large margin on ANLI, RTE, BoolQ, AI2-ARC, OpenbookQA, and StoryCloze. Ablation studies reveal that number of tasks and model scale are key components to the success of instruction tuning.

Finetuned Language Models Are Zero-Shot Learners

Humans (e.g., crowdworkers) have a remarkable ability in solving different tasks, by simply reading textual instructions that define them and looking at a few examples. NLP models built with the conventional paradigm, however, often struggle with generalization across tasks (e.g., a question-answering system cannot solve classification tasks). A long-standing challenge in AI is to build a model that is equipped with the understanding of human-readable instructions that define the tasks, and can generalize to new tasks. To study this, we introduce NATURAL INSTRUCTIONS, a dataset of 61 distinct tasks, their human-authored instructions and 193k task instances. The instructions are obtained from crowdsourcing instructions used to collect existing NLP datasets and mapped to a unified schema. We adopt generative pre-trained language models to encode task-specific instructions along with input and generate task output. Our results indicate that models can benefit from instructions to generalize across tasks. These models, however, are far behind supervised task-specific models, indicating significant room for more progress in this direction.

Cross-Task Generalization via Natural Language Crowdsourcing Instructions

Humans (e.g., crowdworkers) have a remarkable ability in solving different tasks, by simply reading textual instructions that define them and looking at a few examples. Despite the success of the conventional supervised learning on individual datasets, such models often struggle with generalization across tasks (e.g., a question-answering system cannot solve classification tasks). A long-standing challenge in AI is to build a model that learns a new task by understanding the human-readable instructions that define it. To study this, we introduce NATURAL INSTRUCTIONS, a dataset of 61 distinct tasks, their human-authored instructions, and 193k task instances (input-output pairs). The instructions are obtained from crowdsourcing instructions used to create existing NLP datasets and mapped to a unified schema. Using this meta-dataset, we measure cross-task generalization by training models on seen tasks and measuring generalization to the remaining unseen ones. We adopt generative pre-trained language models to encode task-specific instructions along with input and generate task output. Our results indicate that models benefit from instructions when evaluated in terms of generalization to unseen tasks (19% better for models utilizing instructions). These models, however, are far behind an estimated performance upperbound indicating significant room for more progress in this direction. 

/pdf/cross-task-generalization-via-natural-language-crowdsourcing-385vkmgp.pdf

We introduce MetaICL (Meta-training for In-Context Learning), a new meta-training framework for few-shot learning where a pretrained language model is tuned to do in-context learn-ing on a large set of training tasks. This meta-training enables the model to more effectively learn a new task in context at test time, by simply conditioning on a few training examples with no parameter updates or task-specific templates. We experiment on a large, diverse collection of tasks consisting of 142 NLP datasets including classification, question answering, natural language inference, paraphrase detection and more, across seven different meta-training/target splits. MetaICL outperforms a range of baselines including in-context learning without meta-training and multi-task learning followed by zero-shot transfer. We find that the gains are particularly significant for target tasks that have domain shifts from the meta-training tasks, and that using a diverse set of the meta-training tasks is key to improvements. We also show that MetaICL approaches (and sometimes beats) the performance of models fully finetuned on the target task training data, and outperforms much bigger models with nearly 8x parameters.

MetaICL: Learning to Learn In Context

Sewon Min, Mike Lewis, Luke Zettlemoyer, Hannaneh Hajishirzi. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2022. 

https://aclanthology.org/2022.naacl-main.201.pdf

Large language models have recently been shown to attain reasonable zero-shot
generalization on a diverse set of tasks. It has been hypothesized that this is
a consequence of implicit multitask learning in language model training. Can
zero-shot generalization instead be directly induced by explicit multitask
learning? To test this question at scale, we develop a system for easily
mapping general natural language tasks into a human-readable prompted form. We
convert a large set of supervised datasets, each with multiple prompts using
varying natural language. These prompted datasets allow for benchmarking the
ability of a model to perform completely unseen tasks specified in natural
language. We fine-tune a pretrained encoder-decoder model on this multitask
mixture covering a wide variety of tasks. The model attains strong zero-shot
performance on several standard datasets, often outperforming models 16x its
size. Further, our approach attains strong performance on a subset of tasks
from the BIG-Bench benchmark, outperforming models 6x its size. All prompts and
trained models are available at github.com/bigscience-workshop/promptsource/.

Jos Rozen

Papers

Multitask Prompted Training Enables Zero-Shot Task Generalization