# Not All Dialogues are Created Equal: Instance Weighting for Neural Conversational Models

## TL;DR

This paper introduces a weighting model for neural conversational systems that assigns importance scores to training examples, improving performance on dialogue modeling tasks using large, noisy datasets.

## Contribution

It proposes a novel weighting mechanism integrated into neural dialogue models, addressing data quality issues in large unstructured corpora.

## Key findings

- Improved performance on unsupervised dialogue metrics
- Effective handling of noisy, unsegmented dialogue data
- Enhanced model robustness through learned sample weights

## Abstract

Neural conversational models require substantial amounts of dialogue data for their parameter estimation and are therefore usually learned on large corpora such as chat forums or movie subtitles. These corpora are, however, often challenging to work with, notably due to their frequent lack of turn segmentation and the presence of multiple references external to the dialogue itself. This paper shows that these challenges can be mitigated by adding a weighting model into the architecture. The weighting model, which is itself estimated from dialogue data, associates each training example to a numerical weight that reflects its intrinsic quality for dialogue modelling. At training time, these sample weights are included into the empirical loss to be minimised. Evaluation results on retrieval-based models trained on movie and TV subtitles demonstrate that the inclusion of such a weighting model improves the model performance on unsupervised metrics.