Bootstrapping incremental dialogue systems from minimal data: the generalisation power of dialogue grammars

TL;DR

This paper presents an end-to-end approach for inducing task-based dialogue systems from minimal unannotated data, leveraging rich linguistic grammar and reinforcement learning to enable natural, incremental dialogues with high generalization.

Contribution

It introduces a novel combination of incremental semantic grammar (DS-TTR) with reinforcement learning for dialogue system induction from very small datasets, demonstrating strong generalization capabilities.

Findings

The model processes 74% of Facebook AI bAbI dataset with only 0.13% training data.

It processes 65% of bAbI+ with added dialogue phenomena.

Compared to MemN2N, the proposed model shows superior robustness to dialogue variations.

Abstract

We investigate an end-to-end method for automatically inducing task-based dialogue systems from small amounts of unannotated dialogue data. It combines an incremental semantic grammar - Dynamic Syntax and Type Theory with Records (DS-TTR) - with Reinforcement Learning (RL), where language generation and dialogue management are a joint decision problem. The systems thus produced are incremental: dialogues are processed word-by-word, shown previously to be essential in supporting natural, spontaneous dialogue. We hypothesised that the rich linguistic knowledge within the grammar should enable a combinatorially large number of dialogue variations to be processed, even when trained on very few dialogues. Our experiments show that our model can process 74% of the Facebook AI bAbI dataset even when trained on only 0.13% of the data (5 dialogues). It can in addition process 65% of bAbI+, a corpus we created by systematically adding incremental dialogue phenomena such as restarts and self-corrections to bAbI. We compare our model with a state-of-the-art retrieval model, MemN2N. We find that, in terms of semantic accuracy, MemN2N shows very poor robustness to the bAbI+ transformations even when trained on the full bAbI dataset.