Filters








2,880 Hits in 8.2 sec

Improving End-to-End Task-Oriented Dialog System with A Simple Auxiliary Task

Yohan Lee
2021 Findings of the Association for Computational Linguistics: EMNLP 2021   unpublished
Improving End-to-End Task-Oriented Dialogue System with A Simple Auxiliary Task Yohan Lee Electronics  ...  Sequicity: Simplifying Task-oriented Dialogue Systems with Single Sequence-to-Sequence Architectures.  ... 
doi:10.18653/v1/2021.findings-emnlp.112 fatcat:jorqcz2cjrbhvfneqg2tohxcxq

Learning to Learn End-to-End Goal-Oriented Dialog From Related Dialog Tasks [article]

Janarthanan Rajendran, Jonathan K. Kummerfeld, Satinder Singh
2021 arXiv   pre-print
For each goal-oriented dialog task of interest, large amounts of data need to be collected for end-to-end learning of a neural dialog system.  ...  We describe a meta-learning based method that selectively learns from the related dialog task data. Our approach leads to significant accuracy improvements in an example dialog task.  ...  The bAbI dialog tasks are a testbed to evaluate the strengths and shortcomings of end-to-end dialog systems in goal-oriented applications.  ... 
arXiv:2110.15724v1 fatcat:lu72zkqshrexhbsmoh7gr5u2ai

HarperValleyBank: A Domain-Specific Spoken Dialog Corpus [article]

Mike Wu, Jonathan Nafziger, Anthony Scodary, Andrew Maas
2021 arXiv   pre-print
The data size and domain specificity makes for quick transcription experiments with modern end-to-end neural approaches.  ...  We introduce HarperValleyBank, a free, public domain spoken dialog corpus.  ...  The dataset is representative of human to human goal-oriented dialogs for consumer banking with a narrowly scoped set of intents.  ... 
arXiv:2010.13929v2 fatcat:fnxbnyijnvaslakcrns3xi5izm

A Review of the Research on Dialogue Management of Task-Oriented Systems

Yin Jiang Zhao, Yan Ling Li, Min Lin
2019 Journal of Physics, Conference Series  
Finally, it looks forward to the future research direction of task-oriented dialogue system combined with dialogue management.  ...  Scholars pay more and more attention to the spoken dialogue system after the emergence of deep learning technology.  ...  Deep reinforcement learning technology can greatly improve the task success rate of end-to-end task-based dialog system.  ... 
doi:10.1088/1742-6596/1267/1/012025 fatcat:pjm2wqf77rchnmlrovl3lyuxy4

End-to-End Task-Oriented Dialog Modeling with Semi-Structured Knowledge Management [article]

Silin Gao, Ryuichi Takanobu, Antoine Bosselut, Minlie Huang
2022 arXiv   pre-print
Current task-oriented dialog (TOD) systems mostly manage structured knowledge (e.g. databases and tables) to guide the goal-oriented conversations.  ...  To address this task, we propose a TOD system with semi-structured knowledge management, SeKnow, which extends the belief state to manage knowledge with both structured and unstructured contents.  ...  We would also like to thank the anonymous reviewers for their invaluable suggestions and feedback.  ... 
arXiv:2106.11796v3 fatcat:oxcxbecofvf2baxb6c5uz7yjfe

Constraint based Knowledge Base Distillation in End-to-End Task Oriented Dialogs [article]

Dinesh Raghu, Atishya Jain, Mausam, Sachindra Joshi
2021 arXiv   pre-print
End-to-End task-oriented dialogue systems generate responses based on dialog history and an accompanying knowledge base (KB).  ...  Experimental results on three publicly available task-oriented dialog datasets show that our proposed approach outperforms existing state-of-the-art models.  ...  Acknowledgments This work is supported by IBM AI Horizons Network grant, an IBM SUR award, grants by Google, Bloomberg and 1MG, a Visvesvaraya faculty award by Govt. of India, and the Jai Gupta chair fellowship  ... 
arXiv:2109.07396v1 fatcat:tt4s434oeretbi5mor3hkju3qe

Integrating Pretrained Language Model for Dialogue Policy Learning [article]

Hongru Wang, Huimin Wang, Zezhong Wang, Kam-Fai Wong
2021 arXiv   pre-print
However, the reward can be very sparse for it is usually only provided at the end of a dialog session, which causes unaffordable interaction requirements for an acceptable dialog agent.  ...  training into two steps: 1) we integrate a pre-trained language model as a discriminator to judge whether the current system action is good enough for the last user action (i.e., next action prediction  ...  In this case, other components in task-oriented dialogue system are BERTNLU, RuleDST and Template-based NLG.  ... 
arXiv:2111.01398v1 fatcat:4rph6zz5dfeidp7jz5qkymeb5e

CoDraw: Collaborative Drawing as a Testbed for Grounded Goal-driven Communication

Jin-Hwa Kim, Nikita Kitaev, Xinlei Chen, Marcus Rohrbach, Byoung-Tak Zhang, Yuandong Tian, Dhruv Batra, Devi Parikh
2019 Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics  
We present models for our task and benchmark them using both fully automated evaluation and by having them play the game live with humans.  ...  The two players communicate with each other using natural language. We collect the CoDraw dataset of ∼10K dialogs consisting of ∼138K messages exchanged between human players.  ...  End-to-end goal-driven dialog.  ... 
doi:10.18653/v1/p19-1651 dblp:conf/acl/KimKCRZTBP19 fatcat:tl2ixoqdh5bt7bj3ftdj4rh6fe

AidIR: An Interactive Dialog System to Aid Disease Information Retrieval

Da-Jinn Wang, Tsong-Yi Chen, Chia-Yi Su
2022 Applied Sciences  
Finally, we trained the dialog policy network with supervised learning tasks and deployed the reinforcement learning algorithm to allow AidIR to continue learning the dialog policy.  ...  This paper proposes an interactive dialog system, called AidIR, to aid information retrieval.  ...  Task-Oriented Dialog System Task-oriented dialog systems are the backbone of AidIR.  ... 
doi:10.3390/app12041875 fatcat:sjj4jbvlf5dc3n3z42ujgfsrhe

A Survey of Pretrained Language Models Based Text Generation [article]

Junyi Li, Tianyi Tang, Wayne Xin Zhao, Jian-Yun Nie, Ji-Rong Wen
2022 arXiv   pre-print
Next, we present a summary of various useful resources and typical text generation applications to work with PLMs.  ...  In this survey, we begin with introducing three key aspects of applying PLMs to text generation: 1) how to encode the input as representations preserving input semantics which can be fused into PLMs; 2  ...  SC-GPT [143] serializes the system action as the input of and generates according response. Moreover, researchers also attempt to build an end-to-end system for task-oriented dialog.  ... 
arXiv:2201.05273v3 fatcat:nsobhudjwvhwfftrqwklmbiai4

Multi-domain spoken language understanding with transfer learning

Minwoo Jeong, Gary Geunbae Lee
2009 Speech Communication  
To implement multi-domain SLU with transfer learning, we introduce a triangular-chain structured model.  ...  We present a transfer learning approach to the multi-domain SLU problem in which multiple domain-specific data sources can be incorporated.  ...  We would also like to thank Donghyun Lee for his preparation of speech recognition results, and Derek Lactin for his proof-reading of the paper.  ... 
doi:10.1016/j.specom.2009.01.001 fatcat:u4yhnffhfbhwbeg2mgezszcz4u

CoDraw: Collaborative Drawing as a Testbed for Grounded Goal-driven Communication [article]

Jin-Hwa Kim, Nikita Kitaev, Xinlei Chen, Marcus Rohrbach, Byoung-Tak Zhang, Yuandong Tian, Dhruv Batra, Devi Parikh
2019 arXiv   pre-print
We present models for our task and benchmark them using both fully automated evaluation and by having them play the game live with humans.  ...  The two players communicate with each other using natural language. We collect the CoDraw dataset of ~10K dialogs consisting of ~138K messages exchanged between human players.  ...  End-to-end goal-driven dialog.  ... 
arXiv:1712.05558v3 fatcat:mxgfecfzhvdcliqqwicnmvdnf4

DG2: Data Augmentation Through Document Grounded Dialogue Generation [article]

Qingyang Wu, Song Feng, Derek Chen, Sachindra Joshi, Luis A. Lastras, Zhou Yu
2021 arXiv   pre-print
Especially in document-grounded dialog systems, human experts need to carefully read the unstructured documents to answer the users' questions.  ...  Collecting data for training dialog systems can be extremely expensive due to the involvement of human participants and need for extensive annotation.  ...  It has been widely used in many NLP tasks including natural language understanding, question answering, and task-oriented dialog systems to improve the downstream models' performance.  ... 
arXiv:2112.08342v1 fatcat:57dvl3aju5aavanpjqjusfnkhq

Meta-Learning for Low-resource Natural Language Generation in Task-oriented Dialogue Systems [article]

Fei Mi, Minlie Huang, Jiyong Zhang, Boi Faltings
2019 arXiv   pre-print
Natural language generation (NLG) is an essential component of task-oriented dialogue systems.  ...  Meta-NLG defines a set of meta tasks, and directly incorporates the objective of adapting to new low-resource NLG tasks into the meta-learning optimization process.  ...  It is also important and attractive for a task-oriented dialog system to adapt to new functions, namely, supporting new dialog acts that the system has never observed before.  ... 
arXiv:1905.05644v1 fatcat:jv4gkacrevbm7dal3sxycwfpmm

Rethinking Action Spaces for Reinforcement Learning in End-to-end Dialog Agents with Latent Variable Models [article]

Tiancheng Zhao, Kaige Xie, Maxine Eskenazi
2019 arXiv   pre-print
This paper proposes a novel latent action framework that treats the action spaces of an end-to-end dialog agent as latent variables and develops unsupervised methods in order to induce its own action space  ...  Results show that the proposed latent actions achieve superior empirical performance improvement over previous word-level policy gradient methods on both DealOrNoDeal and MultiWoz dialogs.  ...  Conversely, end-to-end (E2E) dialog systems have removed this limit by directly learning a response generation model conditioned on the dialog context using neural networks (Vinyals and Le, 2015; Sordoni  ... 
arXiv:1902.08858v2 fatcat:hkiixigjlzbq7njrmw3ztoohpa
« Previous Showing results 1 — 15 out of 2,880 results