Another Diversity-Promoting Objective Function for Neural Dialogue Generation

Understand

Although generation-based dialogue systems have been widely researched, the response generations by most existing systems have very low diversities.

• The most likely reason for this problem is Maximum Likelihood Estimation (MLE) with Softmax Cross-Entropy (SCE) loss.
• MLE trains models to generate the most frequent responses from enormous generation candidates, although in actual dialogues there are various responses based on the context.
• In this paper, we propose a new objective function called Inverse Token Frequency (ITF) loss, which individually scales smaller loss for frequent token classes and larger loss for rare token classes.