Balancing Multi-Domain Corpora Learning for Open-Domain Response Generation
Yujie Xing, Jinglun Cai, Nils Barlaug, Peng Liu, Jon Atle Gulla · Findings of the Association for Computational Linguistics: NAACL 2022 · 2022
Open-domain conversational systems are assumed to generate equally good responses on multiple domains.Previous work achieved good performance on the single corpus, but training and evaluating on multiple corpora from different domains are less studied.This paper explores methods of generating relevant responses for each of multiple multi-domain corpora.We first examine interleaved learning which intermingles multiple corpora as the baseline.We then investigate two multidomain learning methods, labeled learning and multi-task labeled learning, which encode each corpus through a unique corpus embedding.Furthermore, we propose Domainspecific Frequency (DF), a novel word-level importance weight that measures the relative importance of a word for a specific corpus compared to other corpora.Based on DF, we propose weighted learning, a method that integrates DF to the loss function.We also adopt DF as a new evaluation metric.Extensive experiments show that our methods gain significant improvements on both automatic and human evaluation.We share our code and data for reproducibility. 1 * This work was done prior to the author joining Amazon. 1 https://github.com/yujie-xing/ Balancing_Multi_Domain_Corpus_Learning_ for_Open_Domain_Response_Generation Context What are you going to do on the remote system exactly?PersonaChat I am going to be a pilot.I am going to fly planes.4 corpora (concatenated) I am going to go to the beach.