scholarly journals HyKnow: End-to-End Task-Oriented Dialog Modeling with Hybrid Knowledge Management

Author(s):  
Silin Gao ◽  
Ryuichi Takanobu ◽  
Wei Peng ◽  
Qun Liu ◽  
Minlie Huang
2020 ◽  
Author(s):  
Teakgyu Hong ◽  
Oh-Woog Kwon ◽  
Young-Kil Kim

Author(s):  
Bowen Zhang ◽  
Xiaofei Xu ◽  
Xutao Li ◽  
Yunming Ye ◽  
Xiaojun Chen ◽  
...  
Keyword(s):  

Author(s):  
Florian Strub ◽  
Harm de Vries ◽  
Jérémie Mary ◽  
Bilal Piot ◽  
Aaron Courville ◽  
...  

End-to-end design of dialogue systems has recently become a popular research topic thanks to powerful tools such as encoder-decoder architectures for sequence-to-sequence learning. Yet, most current approaches cast human-machine dialogue management as a supervised learning problem, aiming at predicting the next utterance of a participant given the full history of the dialogue. This vision may fail to correctly render the planning problem inherent to dialogue as well as its contextual and grounded nature. In this paper, we introduce a Deep Reinforcement Learning method to optimize visually grounded task-oriented dialogues, based on the policy gradient algorithm. This approach is tested on the question generation task from the dataset GuessWhat?! containing 120k dialogues and provides encouraging results at solving both the problem of generating natural dialogues and the task of discovering a specific object in a complex picture.


2019 ◽  
Vol 23 (3) ◽  
pp. 1989-2002 ◽  
Author(s):  
Haotian Xu ◽  
Haiyun Peng ◽  
Haoran Xie ◽  
Erik Cambria ◽  
Liuyang Zhou ◽  
...  

2011 ◽  
Vol 36 (3) ◽  
pp. 21-46
Author(s):  
Arijit Laha

In an ideal Knowledge Management environment in an organization, two objectives need to be achieved. Firstly, knowledge workers should have customized informational support for their respective works and secondly, workers across the organization should be able to easily understand and utilize information produced from myriads of knowledge works. Unfortunately, in current KM research and practices, these two goals are rarely addressed together. In fact, most of the KM practices subscribe either to the task-based KM approaches or to the generic/universalistic KM approaches. Typically, each of them is either unable to cater to the need of the other category or provide some ad hoc measures. This paper examines the major issues from a very basic level to understand the problems and attempts to present a solution that systematically covers both the objectives of KM. In the process, it develops a theory, the Task-oriented Organizational Knowledge Management (TOKM), within which the problems are analysed and a viable solution is identified. TOKM gives us a set of design principles for building a new class of IT-based support systems which can serve as a major component of organizational KM. TOKM focuses on information usage in knowledge works and the scope of technology intervention in the related processes. In this paper, the Task-oriented Organizational Knowledge Management is presented as an Information System Design Theory (ISDT) for building integrated IT platforms for supporting organizational KM. In developing the design, the information requirements of knowledge workers in light of an information usage model of knowledge works is studied. Then the model is extended to study possibilities of more advanced IT support and formulate them in the form of a set of meta-requirements. Following the IS design theory paradigm, a set of artifacts are hypothesized to meet the requirements. Finally, a design method, as a possible approach of building an IT-based integrated platform, the Knowledge-work Support Platform (KwSP), is outlined to realize the artifacts in order to meet the requirements. KwSP is a powerful platform for building and maintaining a number of task-type specific Knowledge-work Support Systems (KwSS) on a common sharable platform. Each KwSS, for the task-type supported by it, can be easily designed to provide extensive and sophisticated support to individual as well as group of knowledge workers in performing their respective knowledge work instances.


Author(s):  
Libo Qin ◽  
Xiao Xu ◽  
Wanxiang Che ◽  
Yue Zhang ◽  
Ting Liu
Keyword(s):  

2021 ◽  
Author(s):  
Yanjie Gou ◽  
Yinjie Lei ◽  
Lingqiao Liu ◽  
Yong Dai ◽  
Chunxu Shen

2018 ◽  
Author(s):  
Bing Liu ◽  
Gokhan Tür ◽  
Dilek Hakkani-Tür ◽  
Pararth Shah ◽  
Larry Heck

Sign in / Sign up

Export Citation Format

Share Document