Topic-Oriented Spoken Dialogue Summarization for Customer Service with Saliency-Aware Topic Modeling

  title={Topic-Oriented Spoken Dialogue Summarization for Customer Service with Saliency-Aware Topic Modeling},
  author={Yicheng Zou and Lujun Zhao and Yangyang Kang and Jun Lin and Minlong Peng and Zhuoren Jiang and Changlong Sun and Qi Zhang and Xuanjing Huang and Xiaozhong Liu},
In a customer service system, dialogue summarization can boost service efficiency by automatically creating summaries for long spoken dialogues in which customers and agents try to address issues about specific topics. In this work, we focus on topic-oriented dialogue summarization, which generates highly abstractive summaries that preserve the main ideas from dialogues. In spoken dialogues, abundant dialogue noise and common semantics could obscure the underlying informative content, making… 

Figures and Tables from this paper

CSDS: A Fine-Grained Chinese Dataset for Customer Service Dialogue Summarization

This paper presents a new generation of probabilistic models that combine reinforcement learning with traditional reinforcement learning techniques and shows real-time improvements in the quality and efficiency of these systems.

Topic Model on Microblog with Dual-Streams Graph Convolution Networks

A Dual-streams Graph convolution networks based Topic Model (DGTM) for microblog conversations aims to learn the contexts with full spread mode feature to alleviate data sparsity of microblog posts.

ADPL: Adversarial Prompt-based Domain Adaptation for Dialogue Summarization with Knowledge Disentanglement

An efficient Adversarial Disentangled Prompt Learning (ADPL) model for domain adaptation in dialogue summarization and three kinds of prompts including domain-invariant prompt, domain-specific prompt, and task-oriented prompt are introduced.

Dialogue Topic Segmentation via Parallel Extraction Network with Neighbor Smoothing

This work proposes the parallel extraction network to perform segment extractions, optimizing the bipartite matching cost of segments to capture inter-segment dependency, and proposes neighbor smoothing to handle the segment-bound noise and ambiguity.

GATSum: Graph-Based Topic-Aware Abstract Text Summarization

The result shows that GTASum is superior to many extractive and abstract approaches in terms of ROUGE measurement and proves that the model has the ability to capture the original subject and the correct information and improve the factual accuracy of the summarization.

Domain-Oriented Prefix-Tuning: Towards Efficient and Generalizable Fine-tuning for Zero-Shot Dialogue Summarization

An efficient and generalizable Domain-Oriented Prefix-tuning model, which utilizes a domain word initialized prefix module to alleviate domain entanglement and adopts discrete prompts to guide the model to focus on key contents of dialogues and enhance model generalization is proposed.

Other Roles Matter! Enhancing Role-Oriented Dialogue Summarization via Role Interactions

Role-oriented dialogue summarization is to generate summaries for different roles in the dialogue, e.g., merchants and consumers. Existing methods handle this task by summarizing each role’s content

Heuristic-based Inter-training to Improve Few-shot Multi-perspective Dialog Summarization

This work study the multi-perspective summarization of customer-care conversations between support agents and customers finds that there are different heuristics that are associated with summaries of different perspectives, and explores these heuristic to cre-ate weak-labeled data for intermediate training of the models before intermediate training with scarce human annotated summaries.

DialogLM: Pre-trained Model for Long Dialogue Understanding and Summarization

This work proposes a window-based denoising approach for generative pre-training for long dialogue understanding and summarization, and shows that the pre-trained model DialogLM significantly surpasses the state-of-the-art models across datasets and tasks.



Bleu: a Method for Automatic Evaluation of Machine Translation

This work proposes a method of automatic machine translation evaluation that is quick, inexpensive, and language-independent, that correlates highly with human evaluation, and that has little marginal cost per run.

Attention is All you Need

A new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely is proposed, which generalizes well to other tasks by applying it successfully to English constituency parsing both with large and limited training data.

Contrastive Attention Mechanism for Abstractive Sentence Summarization

The proposed contrastive attention mechanism is more focused on the relevant parts for the summary than the conventional attention mechanism, and greatly advances the state-of-the-art performance on the abstractive sentence summarization task.

Automatic Dialogue Summary Generation for Customer Service

This paper introduces auxiliary key point sequences to solve the problem of dialogue summarization and proposes a novel Leader-Writer network that outperforms other models not only on BLEU and ROUGE-L score but also on logic and integrity.

HIBERT: Document Level Pre-training of Hierarchical Bidirectional Transformers for Document Summarization

This work proposes Hibert (as shorthand for HIerachical Bidirectional Encoder Representations from Transformers) for document encoding and a method to pre-train it using unlabeled data and achieves the state-of-the-art performance on these two datasets.

Investigation of Modeling Units for Mandarin Speech Recognition Using Dfsmn-ctc-smbr

It is found that the proposed hybrid Character-Syllable modeling units is the best choice for CTC based acoustic modeling for Mandarin speech recognition in this work since it can dramatically reduce substitution errors in recognition results.

Unsupervised Abstractive Meeting Summarization with Multi-Sentence Compression and Budgeted Submodular Maximization

A novel graph-based framework for abstractive meeting speech summarization that is fully unsupervised and does not rely on any annotations to take exterior semantic knowledge into account, and to design custom diversity and informativeness measures.

Fast Abstractive Summarization with Reinforce-Selected Sentence Rewriting

An accurate and fast summarization model that first selects salient sentences and then rewrites them abstractively to generate a concise overall summary is proposed, which achieves the new state-of-the-art on all metrics on the CNN/Daily Mail dataset, as well as significantly higher abstractiveness scores.

Legal Summarization for Multi-role Debate Dialogue via Controversy Focus Mining and Multi-task Learning

An innovative end-to-end model is proposed that projects the multi-role debate into the controversy focus space, which enables high-quality essential utterance(s) extraction in terms of legal knowledge and judicial factors.

Sentiment Classification in Customer Service Dialogue with Topic-Aware Multi-Task Learning

This study proposes a topic-aware multi-task learning (TML) approach which learns topic-enriched utterance representations in customer service dialogue by capturing various kinds of topic information.