Towards End-to-End Open Conversational Machine Reading

TL;DR

This paper introduces an end-to-end unified approach for open-retrieval conversational machine reading, improving over traditional cascaded methods by reducing error propagation and achieving state-of-the-art results.

Contribution

It proposes a novel unified text-to-text framework for OR-CMR, replacing separate modules with an end-to-end model for better optimization and performance.

Findings

Significant performance improvements on ShARC and OR-ShARC datasets.

Achieved new state-of-the-art results in OR-CMR tasks.

Framework generalizes across different backbone models.

Abstract

In open-retrieval conversational machine reading (OR-CMR) task, machines are required to do multi-turn question answering given dialogue history and a textual knowledge base. Existing works generally utilize two independent modules to approach this problem's two successive sub-tasks: first with a hard-label decision making and second with a question generation aided by various entailment reasoning methods. Such usual cascaded modeling is vulnerable to error propagation and prevents the two sub-tasks from being consistently optimized. In this work, we instead model OR-CMR as a unified text-to-text task in a fully end-to-end style. Experiments on the ShARC and OR-ShARC dataset show the effectiveness of our proposed end-to-end framework on both sub-tasks by a large margin, achieving new state-of-the-art results. Further ablation studies support that our framework can generalize to different backbone models.