A Diversity-Promoting Objective Function for Neural Conversation Models

Understand

Sequence-to-sequence neural network models for generation of conversational responses tend to generate safe, commonplace responses (e.g., "I don't know") regardless of the input.

• We suggest that the traditional objective function, i.e., the likelihood of output (response) given input (message) is unsuited to response generation tasks.
• Instead we propose using Maximum Mutual Information (MMI) as the objective function in neural models.
• Experimental results demonstrate that the proposed MMI models produce more diverse, interesting, and appropriate responses, yielding substantive gains in BLEU scores on two conversational datasets and in human evaluations.