CLUES: A Benchmark for Learning Classifiers using Natural Language Explanations

TL;DR

This paper introduces CLUES, a benchmark for training zero-shot classifiers from natural language explanations, and proposes ExEnt, an entailment-based model that leverages explanations to improve classification on structured data.

Contribution

The paper presents CLUES, a new benchmark for learning classifiers from explanations, and introduces ExEnt, a model that effectively uses explanations to enhance zero-shot classification performance.

Findings

ExEnt outperforms baseline models by up to 18% on novel tasks.

CLUES includes 180 classification tasks with diverse explanations.

Using explanations improves zero-shot classification accuracy.

Abstract

Supervised learning has traditionally focused on inductive learning by observing labeled examples of a task. In contrast, humans have the ability to learn new concepts from language. Here, we explore training zero-shot classifiers for structured data purely from language. For this, we introduce CLUES, a benchmark for Classifier Learning Using natural language ExplanationS, consisting of a range of classification tasks over structured data along with natural language supervision in the form of explanations. CLUES consists of 36 real-world and 144 synthetic classification tasks. It contains crowdsourced explanations describing real-world tasks from multiple teachers and programmatically generated explanations for the synthetic tasks. To model the influence of explanations in classifying an example, we develop ExEnt, an entailment-based model that learns classifiers using explanations. ExEnt generalizes up to 18% better (relative) on novel tasks than a baseline that does not use explanations. We delineate key challenges for automated learning from explanations, addressing which can lead to progress on CLUES in the future. Code and datasets are available at: https://clues-benchmark.github.io.