EMNLP2020 | 近期必读Dialogue Generation精选论文

AMiner平台由清华大学计算机系研发，拥有我国完全自主知识产权。平台包含了超过2.3亿学术论文/专利和1.36亿学者的科技图谱，提供学者评价、专家发现、智能指派、学术地图等科技情报专业化服务。系统2006年上线，吸引了全球220个国家/地区1000多万独立IP访问，数据下载量230万次，年度访问量超过1100万，成为学术搜索和社会网络挖掘研究的重要数据和实验平台。

AMiner平台：https://www.aminer.cn

导语：EMNLP，自然语言处理经验方法会议（Conference on Empirical Methods in Natural Language Processing），是由国际语言学会（ACL）下属的SIGDAT小组主办的自然语言处理领域的*国际会议，也是自然语言算法的A类会议。EMNLP2020共审阅论文3359篇，接收754篇，接收率为22.4%。

对话系统或对话代理（Dialogue system）是旨在与人对话的计算机系统。对话系统采用文本，语音，图形，触觉，手势和其他模式中的一种或多种在输入和输出通道上进行通信。对话系统目前是自然语言处理中难度较大的研究课题，其元素尚未定义，但是它们与闲聊机器人不同，而是更倾向于以任务为导向，通过对话的方式帮助用户完成具体的任务。

根据AMiner-EMNLP2020词云图和论文可以看出，Dialogue Generation在本次会议中也有许多不凡的工作，下面我们一起看看Dialogue Generation主题的相关论文。

1.论文名称：UniConv: A Unified Conversational Neural Architecture for Multi-domain Task-oriented Dialogues

论文链接：https://www.aminer.cn/pub/5eaaa1d691e011fa9e15eae1?conf=emnlp2020

作者：Le Hung, Sahoo Doyen, Liu Chenghao, Chen Nancy F., Hoi Steven C. H.

简介：

This traditional pipeline modular framework has achieved remarkable successes in task-oriented dialogues. Such kind of dialogue system is not fully optimized as the modules are loosely integrated and often not trained jointly in an end-to-end manner, and may suffer from increasing error propagation between the modules as the complexity of the dialogues evolves.
The authors proposed UniConv, a novel unified neural architecture of conversational agents for Multi-domain Task-oriented Dialogues.
The model jointly trains (1) a Bi-level State Tracker to capture dependencies in both domain and slot levels simultaneously, and (2) a Joint Dialogue Act and Response Generator to model dialogue act latent variable and semantically conditions output responses with contextual cues.
The promising performance of UniConv on the MultiWOZ benchmark validates the efficacy of the method.

2.论文名称：ToD-BERT: Pre-trained Natural Language Understanding for Task-Oriented Dialogues

论文链接：https://www.aminer.cn/pub/5e982ccd91e0119e8a9524cb?conf=emnlp2020

作者：Wu Chien-Sheng, Hoi Steven, Socher Richard, Xiong Caiming

简介：

Recent advances in pre-training using self-attention encoder architectures have been commonly used in many NLP applications.
The authors propose task-oriented dialogue BERT (ToDBERT) that is trained on nine English-based, human-human, multi-turn and publicly available task-oriented datasets across over 60 domains.
ToD-BERT outperforms BERT on four dialogue downstream tasks, including intention classification, dialogue state tracking, dialogue act prediction, and response selection.
It has clear advantage in the few-shot experiments than limited labeled data is available.

3.论文名称：Parallel Interactive Networks for Multi-Domain Dialogue State Generation

论文链接：https://www.aminer.cn/pub/5f6332d491e011242e3f2bb4/?conf=emnlp2020

作者：Junfan Chen, Richong Zhang, Yongyi Mao, Jie Xu

简介：

This paper studies the problem of state generation for multi-domain dialogues.
Existing generation based models fail to model the dialogue dependencies and ignore the slot-overlapping problem in MDST.
To overcome the limitation of existing models, the authors present novel Parallel Interactive Networks (PIN) for more accurate and robust dialogue state generation.
The design of the PIN model is inspired by the interactive nature of the dialogues and the overlapping slots in the ontology.
The slot-overlapping problem are solved by introducing the slot-level context.
Empirical studies on two benchmark datasets demonstrate the effectiveness of the PIN model.

4.论文名称：Dialogue Response Ranking Training with Large-Scale Human Feedback Data

论文链接：https://www.aminer.cn/pub/5f61e3aa91e011fae8fd6a87?conf=emnlp2020

作者：Xiang Gao, Yizhe Zhang, Michel Galley, Chris Brockett, Bill Dolan

简介：

Conversing freely in natural language is one of the greatest challenges of artificial intelligence.
The authors leverage Reddit human feedback data to build and release a large-scale training dataset for feedback prediction.
The authors trained GPT-2 based models on 133M pairs of human feedback data and demonstrate that these models outperform several standard baselines.
The conventional dialog perplexity baseline shows little predictive power on Reddit human feedback data.
The authors ensemble the feedback prediction models and a humanlike scoring model to rank the machine generated dialog responses.

5.论文名称：Structured Attention for Unsupervised Dialogue Structure Induction

论文链接：https://www.aminer.cn/pub/5f68707691e011c23f13b4d7?conf=emnlp2020

作者：Liang Qiu, Yizhou Zhao, Weiyan Shi, Yuan Liang, Feng Shi, Tao Yuan, Zhou Yu, Song-Chun Zhu

简介：

It remains an unsolved question for formally evaluating the performance of dialogue structure induction algorithms.
This paper proposed to inject structured attention into variational recurrent neural network models for unsupervised dialogue structure learning.
The authors explored two different structure inductive biases: linear CRF for utterance-level semantic structure induction in two-party dialogues; and non-projective dependency tree for interactive structure learning in multi-party dialogues.
Both models are proved to have a better structure learning performance over the state-of-the-art algorithms.

6.论文名称：BiST: Bi-directional Spatio-Temporal Reasoning for Video-Grounded Dialogues

论文链接：https://www.aminer.cn/pub/5f7fe6d80205f07f68973167?conf=emnlp2020

作者：Hung Le, Doyen Sahoo, Nancy Chen, Steven C.H. Hoi

简介：

A video-grounded dialogue agent aims to converse with humans based on signals from natural language and from other modalities such as sound and vision of the input video.
This is a very complex task as the dialogue agent needs to possess strong language understanding to generate natural responses and sophisticated reasoning over video information, including the related objects, their positions and motions, etc.
The authors proposed BiST, a novel deep neural network approach for video-grounded dialogues and video QA, which exploits the complex visual nuances of videos through a bidirectional reasoning framework in both spatial and temporal dimensions.
The authors’ experimental results show that BiST can extract relevant, high-resolution visual cues from videos and generate quality dialogue responses/answers.

更多EMNLP2020论文，可以关注公众号或者链接直达EMNLP2020专题，最前沿的研究方向和最全面的论文数据等你来~