ContractNLI: A Dataset for Document-level Natural Language Inference for Contracts

TL;DR

This paper introduces ContractNLI, a large dataset for document-level natural language inference on contracts, highlighting the challenges and proposing a strong baseline model that improves evidence identification in lengthy legal documents.

Contribution

The paper presents the first large-scale dataset for contract-level NLI and analyzes the limitations of existing models, proposing a novel span-based evidence identification method.

Findings

Existing models perform poorly on contract NLI.

A new span-based evidence identification approach improves performance.

Contracts' linguistic features like negations increase task difficulty.

Abstract

Reviewing contracts is a time-consuming procedure that incurs large expenses to companies and social inequality to those who cannot afford it. In this work, we propose "document-level natural language inference (NLI) for contracts", a novel, real-world application of NLI that addresses such problems. In this task, a system is given a set of hypotheses (such as "Some obligations of Agreement may survive termination.") and a contract, and it is asked to classify whether each hypothesis is "entailed by", "contradicting to" or "not mentioned by" (neutral to) the contract as well as identifying "evidence" for the decision as spans in the contract. We annotated and release the largest corpus to date consisting of 607 annotated contracts. We then show that existing models fail badly on our task and introduce a strong baseline, which (1) models evidence identification as multi-label classification over spans instead of trying to predict start and end tokens, and (2) employs more sophisticated context segmentation for dealing with long documents. We also show that linguistic characteristics of contracts, such as negations by exceptions, are contributing to the difficulty of this task and that there is much room for improvement.