Jointly Optimizing Diversity and Relevance in Neural Response Generation
This addresses the issue of balancing diversity and relevance in response generation for conversational AI, which is an incremental improvement over existing methods.
The paper tackles the problem of neural conversation models generating bland and generic responses by proposing a SpaceFusion model that jointly optimizes diversity and relevance, resulting in significant improvements in both metrics compared to strong baselines.
Although recent neural conversation models have shown great potential, they often generate bland and generic responses. While various approaches have been explored to diversify the output of the conversation model, the improvement often comes at the cost of decreased relevance. In this paper, we propose a SpaceFusion model to jointly optimize diversity and relevance that essentially fuses the latent space of a sequence-to-sequence model and that of an autoencoder model by leveraging novel regularization terms. As a result, our approach induces a latent space in which the distance and direction from the predicted response vector roughly match the relevance and diversity, respectively. This property also lends itself well to an intuitive visualization of the latent space. Both automatic and human evaluation results demonstrate that the proposed approach brings significant improvement compared to strong baselines in both diversity and relevance.