Controllable Summarization with Constrained Markov Decision Process

TL;DR

This paper introduces a novel framework using Constrained Markov Decision Process for controllable text summarization, enabling explicit control over attributes like length and content while maintaining summary quality.

Contribution

The work presents a new training framework that incorporates constraints and rewards for better attribute control in summarization, improving over existing methods.

Findings

Effective control over summary attributes demonstrated

Summaries remain informative while satisfying constraints

Framework outperforms baseline models on benchmarks

Abstract

We study controllable text summarization which allows users to gain control on a particular attribute (e.g., length limit) of the generated summaries. In this work, we propose a novel training framework based on Constrained Markov Decision Process (CMDP), which conveniently includes a reward function along with a set of constraints, to facilitate better summarization control. The reward function encourages the generation to resemble the human-written reference, while the constraints are used to explicitly prevent the generated summaries from violating user-imposed requirements. Our framework can be applied to control important attributes of summarization, including length, covered entities, and abstractiveness, as we devise specific constraints for each of these aspects. Extensive experiments on popular benchmarks show that our CMDP framework helps generate informative summaries while complying with a given attribute's requirement.