Deep Active Learning for Dialogue Generation
This work addresses the challenge of improving dialogue generation for conversational AI systems, though it appears incremental as it builds on existing active learning and neural methods.
The authors tackled the problem of generating open-domain dialogue by proposing a neural conversational model that combines offline supervised learning with online active learning using one-character user feedback, resulting in semantically relevant and interesting responses and enabling training of agents with customized personas and styles.
We propose an online, end-to-end, neural generative conversational model for open-domain dialogue. It is trained using a unique combination of offline two-phase supervised learning and online human-in-the-loop active learning. While most existing research proposes offline supervision or hand-crafted reward functions for online reinforcement, we devise a novel interactive learning mechanism based on hamming-diverse beam search for response generation and one-character user-feedback at each step. Experiments show that our model inherently promotes the generation of semantically relevant and interesting responses, and can be used to train agents with customized personas, moods and conversational styles.