Affiliation:
1. Institute of Computer Science and Technology, Peking University
2. Information Sciences Institute, University of Southern California
3. Beijing Institute of Big Data Research
Abstract
The availability of abundant conversational data on the Internet brought prosperity to the generation-based open domain conversation systems. In the training of the generation models, existing methods generally treat all the training data equivalently. However, the data crawled from the websites may contain many noises. Blindly training with the noisy data could harm the performance of the final generation model. In this paper, we propose a generation with calibration framework, that allows high- quality data to have more influences on the generation model and reduces the effect of noisy data. Specifically, for each instance in training set, we employ a calibration network to produce a quality score for it, then the score is used for the weighted update of the generation model parameters. Experiments show that the calibrated model outperforms baseline methods on both automatic evaluation metrics and human annotations.
Publisher
International Joint Conferences on Artificial Intelligence Organization
Cited by
9 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献
1. Multi-Turn Dialogue Agent as Sales' Assistant in Telemarketing;2023 International Joint Conference on Neural Networks (IJCNN);2023-06-18
2. Research on Applying the “Shift” Concept to Deep Attention Matching;Applied Sciences;2023-03-20
3. Research on Shift Matching to Enhance DAM;2021 8th International Conference on Information, Cybernetics, and Computational Social Systems (ICCSS);2021-12-10
4. Dynamic Curriculum Learning with Co-training for Medical Dialogue Generation;2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM);2021-12-09
5. Multi-level Spatio-temporal Matching Network for Multi-turn Response Selection in Retrieval-based Dialogue Systems;2021 IEEE International Conference on Big Knowledge (ICBK);2021-12