Multi-task learning with graph attention networks for multi-domain task-oriented dialogue systems

Zeng, Meng; Wang, Lifang; Jiang, Zejun; Li, Ronghan; Lu, Xinyu; Hu, Zhongtian

doi:10.1016/j.knosys.2022.110069

Cited by 20 publications

(4 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Initially, rule-based systems were among the earliest approaches to textual dialogue response generation. Zhao et al [16] introduced a context-aware dialogue response generation method based on heterogeneous attention networks. This approach considers multiple levels of contextual information to achieve highly accurate results.…”

Section: Related Work 21 Textual Dialogue Response Generationmentioning

confidence: 99%

Multimodal Dialogue Response Generation Based on Selective Attention and Gating Mechanisms

Yu,

Zhang,

Ding

et al. 2023

Preprint

View full text Add to dashboard Cite

Picture response is a crucial aspect in dialogue systems; however, existing systems predominantly focus on text information, often neglecting the use of image data or relying on large-scale language models, which leads to high training costs and slower generation speeds. To address these challenges, this paper introduces the MDSAGM (Selective Attention and Gating Mechanism)model, which comprises a Transformer-based Text Dialogue Response Generator, Text-to-Image Generation, and SAGM (Selective Attention and Gating Mechanism). The model is core lies in combining Selective Attention and Gating Mechanism to enhance generalization and improve model accuracy. The study further explores the contributions of image and text information in multimodal fusion. Compared to other large-scale training models, the proposed model demonstrates computational efficiency , fewer parameters, and shorter response times, making it more lightweight. This dialogue model facilitates the separation of multimodal dialogue parameters from the overall model, enabling better parameter fitting through pre-training with abundant plain text and text-image data. Extensive experiments show that the proposed method achieves promising results in both automatic and manual evaluations, generating information-rich text and image responses with higher accuracy while maintaining a lightweight structure.

show abstract

Section: Related Work 21 Textual Dialogue Response Generationmentioning

confidence: 99%

Multimodal Dialogue Response Generation Based on Selective Attention and Gating Mechanisms

Yu,

Zhang,

Ding

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Advanced techniques such as multi-task learning and graph attention networks have also been introduced into knowledge dialogue system research. Zhao et al [17] put forward a multi-task learning framework built on graph attention networks intended for multi-domain goal-driven conversational systems. The introduction of these techniques helps to improve the generalization capability and performance of dialogue systems in different domains.…”

Section: Related Work 21 Dialogue Systemsmentioning

confidence: 99%

FinKENet: A Novel Financial Knowledge Enhanced Network for Financial Question Matching

Guo,

Liang,

Chen

et al. 2023

Entropy

View full text Add to dashboard Cite

Question matching is the fundamental task in retrieval-based dialogue systems which assesses the similarity between Query and Question. Unfortunately, existing methods focus on improving the accuracy of text similarity in the general domain, without adaptation to the financial domain. Financial question matching has two critical issues: (1) How to accurately model the contextual representation of a financial sentence? (2) How to accurately represent financial key phrases in an utterance? To address these issues, this paper proposes a novel Financial Knowledge Enhanced Network (FinKENet) that significantly injects financial knowledge into contextual text. Specifically, we propose a multi-level encoder to extract both sentence-level features and financial phrase-level features, which can more accurately represent sentences and financial phrases. Furthermore, we propose a financial co-attention adapter to combine sentence features and financial keyword features. Finally, we design a multi-level similarity decoder to calculate the similarity between queries and questions. In addition, a cross-entropy-based loss function is presented for model optimization. Experimental results demonstrate the effectiveness of the proposed method on the Ant Financial question matching dataset. In particular, the Recall score improves from 73.21% to 74.90% (1.69% absolute).

show abstract

“…The dialogue system is an important research direction in artificial intelligence [1,2]. In recent years, thanks to the development of deep learning technology and the improvement of hardware computing power, the research of dialogue systems is also advancing rapidly.…”

Section: Introductionmentioning

confidence: 99%

The Advance of Dialogue System with Neural Network

Ding,

Dong,

Wang

2023

ACE

View full text Add to dashboard Cite

With the rapid reform of artificial intelligence, the dialogue system as an important branch of artificial intelligence has also developed greatly in recent years. However, dialogue systems also face many challenges, and accurate and well-sampled training data in dialogue systems is often not readily available. How to use the relationship extraction model based on meta-learning from a small number of samples to improve the accuracy of the relationship prediction method on few-shot learning training data. In addition, this article also explores methods of data compression, summarizing previous compression methods in different ways and related work. At the same time, the application of data compression in the dialogue chat model is also studied. The comparison before and after data compression is carried out to make the dialogue chat model maintain the original answer quality when the data part is reduced, and how to use knowledge graph and reinforcement learning to improve the answer quality of the dialogue system after data compression.

show abstract

Multi-task learning with graph attention networks for multi-domain task-oriented dialogue systems

Cited by 20 publications

References 20 publications

Multimodal Dialogue Response Generation Based on Selective Attention and Gating Mechanisms

Multimodal Dialogue Response Generation Based on Selective Attention and Gating Mechanisms

FinKENet: A Novel Financial Knowledge Enhanced Network for Financial Question Matching

The Advance of Dialogue System with Neural Network

Contact Info

Product

Resources

About