FewJoint: few-shot learning for joint dialogue understanding
International Journal of Machine Learning and Cybernetics, 1--15, 2022.
Hou, Yutai and Wang, Xinghao and Chen, Cheng and Li, Bohan and Che, Wanxiang and Chen, Zhigang
任务型对话系统
任务型对话系统是指以人机对话形式提供信息或服务的系统，它能够帮助人们完成一些垂直领域的服务，例如查天气、酒店预订和订机票等。 近年来，任务型对话系统的研究主要分为两个流派：流水线任务型对话系统和端到端任务型对话系统。其中流水线任务型对话系统主要由自然语言理解模块、对话状态跟踪模块、对话策略学习模块和自然语言生成模块组成，并需要通过各个子模块协同工作一起生成对话系统回复。而端到端任务型对话系统则可以直接通过一个统一的序列到序列模型生成对话系统的回复。
自然语言理解
自然语言理解的目的是将用户的输入映射到预先根据不同场景定义的语义槽中，通常包括三个任务：领域检测、意图识别和语义槽填充。自然语言理解应尽可能完整、清晰和准确地将用户输入转化为计算机能够理解的形式。举例如下：
用户查询：帮我订一张从北京去上海的飞机票
输出结果：领域 —— 出行 意图 —— 订飞机票 语义槽 —— {出发地：北京，目的地：上海}
传统的自然语言理解系统存在的一些问题，如：
- 对于意图与槽位进行分别识别，未能利用意图与槽位体系关系
- 在多领域的迁移中表现较差
- 在小样本场景下效果较差
本组研究专注于
- 如何更好的利用意图与槽位之间的关系来提升自然语言理解模块性能
- 如何在多领域的迁移中获得较好的效果
- 如何提升小样本场景下自然语言理解模块的性能
- 探索任务型对话系统的新场景，如多意图对话
- 探索多语言任务型对话场景下的效果
对话状态跟踪
对话状态是一种将当前时刻的对话表示为可供系统选择下一时刻动作信息的数据结构，可以看作每个槽值的取值分布情况。对话状态跟踪以当前的动作、前轮的对话状态和相应的系统动作作为输入，输出其对当前对话状态的估计。对话策略的选择依赖于对话状态跟踪估计的对话状态，因此对话状态跟踪至关重要。同时，对话状态跟踪也非常具有挑战性，因为自然语言理解模块的识别不一定是完全正确的。
端到端任务型对话系统
深度学习的飞速发展促进了端到端方法在任务型对话系统的应用，使得端到端的任务型对话系统成为可能。端到端方法将流水线方法中的四个模块用统一的端到端方法代替，根据用户的输入，直接生成相应的回复或响应。端到端方法的输入为当前的对话历史以及支撑对话的知识库，输出为系统回复。本组研究专注于：
- 如何更好的维护和更新输入的知识库，以便在输出中回复正确的知识
- 如何更好的在多领域间迁移
目前，在任务型对话系统研究方向，已发表CCF A/B类论文20余篇，相关研究在工业界系统中落地实现应用。
评测成绩：
论文列表
FewJoint: few-shot learning for joint dialogue understanding
International Journal of Machine Learning and Cybernetics, 1--15, 2022.
Hou, Yutai and Wang, Xinghao and Chen, Cheng and Li, Bohan and Che, Wanxiang and Chen, Zhigang
GL-CLeF: A Global-Local Contrastive Learning Framework for Cross-lingual Spoken Language Understanding
Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2677--2686, 2022
Qin, Libo and Chen, Qiguang and Xie, Tianbao and Li, Qixin and Lou, Jian-Guang and Che, Wanxiang and Kan, Min-Yen
GL-CLeF: A Global-Local Contrastive Learning Framework for Cross-lingual Spoken Language Understanding
Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2677--2686, 2022
Qin, Libo and Chen, Qiguang and Xie, Tianbao and Li, Qixin and Lou, Jian-Guang and Che, Wanxiang and Kan, Min-Yen
Inverse is Better! Fast and Accurate Prompt for Few-shot Slot Tagging
Findings of the Association for Computational Linguistics: ACL 2022, 637--647
Hou, Yutai and Chen, Cheng and Luo, Xianzhen and Li, Bohan and Che, Wanxiang
Inverse is Better! Fast and Accurate Prompt for Few-shot Slot Tagging
Findings of the Association for Computational Linguistics: ACL 2022, 637--647
Hou, Yutai and Chen, Cheng and Luo, Xianzhen and Li, Bohan and Che, Wanxiang
Text Is No More Enough! A Benchmark for Profile-Based Spoken Language Understanding
Proceedings of the AAAI Conference on Artificial Intelligence, 11575--11585, 2022.
Xu, Xiao and Qin, Libo and Chen, Kaiji and Wu, Guoxing and Li, Linlin and Che, Wanxiang
Text Is No More Enough! A Benchmark for Profile-Based Spoken Language Understanding
Proceedings of the AAAI Conference on Artificial Intelligence, 11575--11585, 2022.
Xu, Xiao and Qin, Libo and Chen, Kaiji and Wu, Guoxing and Li, Linlin and Che, Wanxiang
A survey on spoken language understanding: Recent advances and new frontiers
ArXiv preprint, abs/2103.03095, 2021.
Qin, Libo and Xie, Tianbao and Che, Wanxiang and Liu, Ting
A survey on spoken language understanding: Recent advances and new frontiers
ArXiv preprint, abs/2103.03095, 2021.
Qin, Libo and Xie, Tianbao and Che, Wanxiang and Liu, Ting
C2c-genda: Cluster-to-cluster generation for data augmentation of slot filling
Proceedings of the AAAI Conference on Artificial Intelligence, 13027--13035, 2021.
Hou, Yutai and Chen, Sanyuan and Che, Wanxiang and Chen, Cheng and Liu, Ting
C2c-genda: Cluster-to-cluster generation for data augmentation of slot filling
Proceedings of the AAAI Conference on Artificial Intelligence, 13027--13035, 2021.
Hou, Yutai and Chen, Sanyuan and Che, Wanxiang and Chen, Cheng and Liu, Ting
Don't be Contradicted with Anything! CI-ToD: Towards Benchmarking Consistency for Task-oriented Dialogue System
ArXiv preprint, abs/2109.11292, 2021.
Qin, Libo and Xie, Tianbao and Huang, Shijue and Chen, Qiguang and Xu, Xiao and Che, Wanxiang
Don't be Contradicted with Anything! CI-ToD: Towards Benchmarking Consistency for Task-oriented Dialogue System
ArXiv preprint, abs/2109.11292, 2021.
Qin, Libo and Xie, Tianbao and Huang, Shijue and Chen, Qiguang and Xu, Xiao and Che, Wanxiang
Few-shot learning for multi-label intent detection
Proceedings of the AAAI Conference on Artificial Intelligence, 13036--13044, 2021.
Hou, Yutai and Lai, Yongkui and Wu, Yushan and Che, Wanxiang and Liu, Ting
Few-shot learning for multi-label intent detection
Proceedings of the AAAI Conference on Artificial Intelligence, 13036--13044, 2021.
Hou, Yutai and Lai, Yongkui and Wu, Yushan and Che, Wanxiang and Liu, Ting
GL-GIN: Fast and Accurate Non-Autoregressive Model for Joint Multiple Intent Detection and Slot Filling
Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), 178--188, 2021.
Qin, Libo and Wei, Fuxuan and Xie, Tianbao and Xu, Xiao and Che, Wanxiang and Liu, Ting
GL-GIN: Fast and Accurate Non-Autoregressive Model for Joint Multiple Intent Detection and Slot Filling
Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), 178--188, 2021.
Qin, Libo and Wei, Fuxuan and Xie, Tianbao and Xu, Xiao and Che, Wanxiang and Liu, Ting
Knowing where to leverage: Context-aware graph convolutional network with an adaptive fusion layer for contextual spoken language understanding
IEEE/ACM Transactions on Audio, Speech, and Language Processing, 1280--1289, 2021.
Qin, Libo and Che, Wanxiang and Ni, Minheng and Li, Yangming and Liu, Ting
Knowing where to leverage: Context-aware graph convolutional network with an adaptive fusion layer for contextual spoken language understanding
IEEE/ACM Transactions on Audio, Speech, and Language Processing, 1280--1289, 2021.
Qin, Libo and Che, Wanxiang and Ni, Minheng and Li, Yangming and Liu, Ting
Learning to Bridge Metric Spaces: Few-shot Joint Learning of Intent Detection and Slot Filling
Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, 3190--3200, 2021.
Hou, Yutai and Lai, Yongkui and Chen, Cheng and Che, Wanxiang and Liu, Ting
Learning to Bridge Metric Spaces: Few-shot Joint Learning of Intent Detection and Slot Filling
Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, 3190--3200, 2021.
Hou, Yutai and Lai, Yongkui and Chen, Cheng and Che, Wanxiang and Liu, Ting
A co-interactive transformer for joint slot filling and intent detection
ArXiv preprint, abs/2010.03880, 2020.
Qin, Libo and Liu, Tailu and Che, Wanxiang and Kang, Bingbing and Zhao, Sendong and Liu, Ting
A co-interactive transformer for joint slot filling and intent detection
ArXiv preprint, abs/2010.03880, 2020.
Qin, Libo and Liu, Tailu and Che, Wanxiang and Kang, Bingbing and Zhao, Sendong and Liu, Ting
AGIF: An Adaptive Graph-Interactive Framework for Joint Multiple Intent Detection and Slot Filling
Findings of the Association for Computational Linguistics: EMNLP 2020, 1807--1816, 2020.
Qin, Libo and Xu, Xiao and Che, Wanxiang and Liu, Ting
AGIF: An Adaptive Graph-Interactive Framework for Joint Multiple Intent Detection and Slot Filling
Findings of the Association for Computational Linguistics: EMNLP 2020, 1807--1816, 2020.
Qin, Libo and Xu, Xiao and Che, Wanxiang and Liu, Ting
Dcr-net: A deep co-interactive relation network for joint dialog act recognition and sentiment classification
Proceedings of the AAAI conference on artificial intelligence, 8665--8672, 2020.
Qin, Libo and Che, Wanxiang and Li, Yangming and Ni, Mingheng and Liu, Ting
Dcr-net: A deep co-interactive relation network for joint dialog act recognition and sentiment classification
Proceedings of the AAAI conference on artificial intelligence, 8665--8672, 2020.
Qin, Libo and Che, Wanxiang and Li, Yangming and Ni, Mingheng and Liu, Ting
Dynamic Fusion Network for Multi-Domain End-to-end Task-Oriented Dialog
Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 6344--6354, 2020.
Qin, Libo and Xu, Xiao and Che, Wanxiang and Zhang, Yue and Liu, Ting
Dynamic Fusion Network for Multi-Domain End-to-end Task-Oriented Dialog
Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 6344--6354, 2020.
Qin, Libo and Xu, Xiao and Che, Wanxiang and Zhang, Yue and Liu, Ting
Injecting word information with multi-level word adapter for chinese spoken language understanding
ArXiv preprint, abs/2010.03903, 2020.
Teng, Dechuan and Qin, Libo and Che, Wanxiang and Zhao, Sendong and Liu, Ting
Injecting word information with multi-level word adapter for chinese spoken language understanding
ArXiv preprint, abs/2010.03903, 2020.
Teng, Dechuan and Qin, Libo and Che, Wanxiang and Zhao, Sendong and Liu, Ting
Multi-domain spoken language understanding using domain-and task-aware parameterization
ArXiv preprint, abs/2004.14871, 2020.
Qin, Libo and Wei, Fuxuan and Ni, Minheng and Zhang, Yue and Che, Wanxiang and Li, Yangming and Liu, Ting
Multi-domain spoken language understanding using domain-and task-aware parameterization
ArXiv preprint, abs/2004.14871, 2020.
Qin, Libo and Wei, Fuxuan and Ni, Minheng and Zhang, Yue and Che, Wanxiang and Li, Yangming and Liu, Ting
Slot-consistent NLG for Task-oriented Dialogue Systems with Iterative Rectification Network
Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 97--106, 2020.
Li, Yangming and Yao, Kaisheng and Qin, Libo and Che, Wanxiang and Li, Xiaolong and Liu, Ting
Slot-consistent NLG for Task-oriented Dialogue Systems with Iterative Rectification Network
Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 97--106, 2020.
Li, Yangming and Yao, Kaisheng and Qin, Libo and Che, Wanxiang and Li, Xiaolong and Liu, Ting
A corpus-free state2seq user simulator for task-oriented dialogue
ArXiv preprint, abs/1909.04448, 2019.
Hou, Yutai and Fang, Meng and Che, Wanxiang and Liu, Ting
A corpus-free state2seq user simulator for task-oriented dialogue
ArXiv preprint, abs/1909.04448, 2019.
Hou, Yutai and Fang, Meng and Che, Wanxiang and Liu, Ting
A Stack-Propagation Framework with Token-Level Intent Detection for Spoken Language Understanding
Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), 2078--2087, 2019.
Qin, Libo and Che, Wanxiang and Li, Yangming and Wen, Haoyang and Liu, Ting
A Stack-Propagation Framework with Token-Level Intent Detection for Spoken Language Understanding
Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), 2078--2087, 2019.
Qin, Libo and Che, Wanxiang and Li, Yangming and Wen, Haoyang and Liu, Ting
Entity-Consistent End-to-end Task-Oriented Dialogue System with KB Retriever
Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), 133--142, 2019.
Qin, Libo and Liu, Yijia and Che, Wanxiang and Wen, Haoyang and Li, Yangming and Liu, Ting
Entity-Consistent End-to-end Task-Oriented Dialogue System with KB Retriever
Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), 133--142, 2019.
Qin, Libo and Liu, Yijia and Che, Wanxiang and Wen, Haoyang and Li, Yangming and Liu, Ting
Sequence-to-Sequence Data Augmentation for Dialogue Language Understanding
Proceedings of the 27th International Conference on Computational Linguistics, 1234--1245, 2018.
Hou, Yutai and Liu, Yijia and Che, Wanxiang and Liu, Ting
Sequence-to-Sequence Data Augmentation for Dialogue Language Understanding
Proceedings of the 27th International Conference on Computational Linguistics, 1234--1245, 2018.
Hou, Yutai and Liu, Yijia and Che, Wanxiang and Liu, Ting
Sequence-to-Sequence Learning for Task-oriented Dialogue with Dialogue State Representation
Proceedings of the 27th International Conference on Computational Linguistics, 3781--3792, 2018.
Wen, Haoyang and Liu, Yijia and Che, Wanxiang and Qin, Libo and Liu, Ting
Sequence-to-Sequence Learning for Task-oriented Dialogue with Dialogue State Representation
Proceedings of the 27th International Conference on Computational Linguistics, 3781--3792, 2018.
Wen, Haoyang and Liu, Yijia and Che, Wanxiang and Qin, Libo and Liu, Ting