TIAGE: A Benchmark for Topic-Shift Aware Dialog Modeling

TL;DR

TIAGE is a new benchmark for evaluating dialog systems' ability to detect and adapt to topic shifts, highlighting the importance of topic-awareness in natural conversations and revealing current challenges in the field.

Contribution

The paper introduces TIAGE, a novel benchmark with human-annotated topic shifts, and proposes three tasks to advance topic-shift modeling in dialog systems.

Findings

Topic-shift signals improve response generation.

Dialog systems struggle to decide when to change topics.

Further research is needed in topic-shift aware dialog modeling.

Abstract

Human conversations naturally evolve around different topics and fluently move between them. In research on dialog systems, the ability to actively and smoothly transition to new topics is often ignored. In this paper we introduce TIAGE, a new topic-shift aware dialog benchmark constructed utilizing human annotations on topic shifts. Based on TIAGE, we introduce three tasks to investigate different scenarios of topic-shift modeling in dialog settings: topic-shift detection, topic-shift triggered response generation and topic-aware dialog generation. Experiments on these tasks show that the topic-shift signals in TIAGE are useful for topic-shift response generation. On the other hand, dialog systems still struggle to decide when to change topic. This indicates further research is needed in topic-shift aware dialog modeling.