Controllable Factuality in Document-Grounded Dialog Systems Using a Noisy Channel Model

TL;DR

This paper introduces a document-grounded dialog response generation model using a noisy channel approach, enabling control over factuality and fluency, and demonstrating improved factual accuracy over baselines.

Contribution

The work presents a novel noisy channel model for document-grounded dialog, allowing explicit control of factuality versus fluency, and combines it with existing methods for enhanced performance.

Findings

Model achieves higher factuality metrics than baseline.

Scaling factors effectively control factuality-fluency tradeoff.

Combining with CTRL yields further improvements.

Abstract

In this work, we present a model for document-grounded response generation in dialog that is decomposed into two components according to Bayes theorem. One component is a traditional ungrounded response generation model and the other component models the reconstruction of the grounding document based on the dialog context and generated response. We propose different approximate decoding schemes and evaluate our approach on multiple open-domain and task-oriented document-grounded dialog datasets. Our experiments show that the model is more factual in terms of automatic factuality metrics than the baseline model. Furthermore, we outline how introducing scaling factors between the components allows for controlling the tradeoff between factuality and fluency in the model output. Finally, we compare our approach to a recently proposed method to control factuality in grounded dialog, CTRL (arXiv:2107.06963), and show that both approaches can be combined to achieve additional improvements.