Generating Dialogue Responses from a Semantic Latent Space
arXiv:2010.01658
Abstract
Existing open-domain dialogue generation models are usually trained to mimic the gold response in the training set using cross-entropy loss on the vocabulary. However, a good response does not need to resemble the gold response, since there are multiple possible responses to a given prompt. In this work, we hypothesize that the current models are unable to integrate information from multiple semantically similar valid responses of a prompt, resulting in the generation of generic and uninformative responses. To address this issue, we propose an alternative to the end-to-end classification on vocabulary. We learn the pair relationship between the prompts and responses as a regression task on a latent space instead. In our novel dialog generation model, the representations of semantically related sentences are close to each other on the latent space. Human evaluation showed that learning the task on a continuous space can generate responses that are both relevant and informative.
EMNLP 2020
References in corpus (14)
- A Neural Conversational Model
- Effective Approaches to Attention-based Neural Machine Translation
- DailyDialog: A Manually Labelled Multi-turn Dialogue Dataset
- How NOT To Evaluate Your Dialogue System: An Empirical Study of Unsupervised Evaluation Metrics for Dialogue Response Generation
- Adversarial Discriminative Domain Adaptation
- Deep Reinforcement Learning for Dialogue Generation
- TransferTransfo: A Transfer Learning Approach for Neural Network Based Conversational Agents
- Variational Lossy Autoencoder
- Personalizing Dialogue Agents: I have a dog, do you have pets too?
- Hierarchical Neural Story Generation
- Learning Deep Latent Spaces for Multi-Label Classification
- Jointly Optimizing Diversity and Relevance in Neural Response Generation
- Deep Canonically Correlated LSTMs
- An Auto-Encoder Matching Model for Learning Utterance-Level Semantic Dependency in Dialogue Generation