Dialog Policy Learning for Joint Clarification and Active Learning Queries

TL;DR

This paper presents a hierarchical dialog policy that jointly learns clarification and active learning strategies to improve interactive language-based image retrieval, outperforming static policies in an online shopping context.

Contribution

It introduces a novel joint learning approach for dialog policies that combine clarification and active learning, enhancing system adaptability and performance.

Findings

Joint learning outperforms static policies

Improved accuracy in image retrieval tasks

Enhanced system adaptability to new concepts

Abstract

Intelligent systems need to be able to recover from mistakes, resolve uncertainty, and adapt to novel concepts not seen during training. Dialog interaction can enable this by the use of clarifications for correction and resolving uncertainty, and active learning queries to learn new concepts encountered during operation. Prior work on dialog systems has either focused on exclusively learning how to perform clarification/ information seeking, or to perform active learning. In this work, we train a hierarchical dialog policy to jointly perform both clarification and active learning in the context of an interactive language-based image retrieval task motivated by an online shopping application, and demonstrate that jointly learning dialog policies for clarification and active learning is more effective than the use of static dialog policies for one or both of these functions.