Research map / Label-efficient and robust learning / Recognition and long-tail
Topic-Oriented Spoken Dialogue Summarization for Customer Service with Saliency-Aware Topic Modeling
Yicheng Zou, Lujun Zhao, Yangyang Kang, Jun Lin, Minlong Peng, Zhuoren Jiang, Changlong Sun, Qi Zhang, Xuanjing Huang, Xiaozhong Liu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Replace generic topic modeling with SATM word-saliency learning | TDS with BERT · Customer service test set | ROUGE-1: 36.13 -> 36.81 | Improved |
| Replace generic topic modeling with SATM word-saliency learning | TDS with BERT · Customer service test set | BLEU: 10.77 -> 11.24 | Improved |
| Add SATM topic information to the two-stage summarizer | FastRL* · Customer service test set | ROUGE-1: +0.82 | Improved |
| Add SATM topic information to the two-stage summarizer | FastRL* · Customer service test set | ROUGE-2: +0.96 | Improved |
| Add SATM topic information to the two-stage summarizer | FastRL* · Customer service test set | ROUGE-L: +0.75 | Improved |
| Add SATM topic information to the two-stage summarizer | FastRL* · Customer service test set | BLEU: +0.84 | Improved |
| Add NTM topic modeling | FastRL · Customer service test set | ROUGE-1 | Improved |
| Add customer-utterance topic modeling (ablation) | TDS+SATM · Customer service test set | BLEU: 11.09 -> 11.24 | No clear effect |
| Add agent-utterance topic modeling (ablation) | TDS+SATM · Customer service test set | BLEU: 10.73 -> 11.24 | Improved |
| Add at least one role-specific topic model (ablation) | TDS+SATM · Customer service test set | BLEU: 10.37 -> 11.24 | Improved |
| Use more than 50 topics (ablation) | TDS+SATM · Customer service test set | summarization performance | Hurt |
| Use an overly unbalanced ratio of other to total topics (ablation) | TDS+SATM · Customer service test set | summarization performance | Hurt |
Closest papers by meaning
- Sentiment Classification in Customer Service Dialogue with Topic-Aware Multi-Task Learning (AAAI 2020 · similarity 0.87)
- Learning to Align Question and Answer Utterances in Customer Service Conversation with Recurrent Pointer Networks (AAAI 2019 · similarity 0.84)
- icsPLMs: Exploring Pre-trained Language Models in Intelligent Customer Service (Student Abstract) (AAAI 2024 · similarity 0.82)
- Distilling Structured Rationale from Large Language Models to Small Language Models for Abstractive Summarization (AAAI 2025 · similarity 0.80)
- Mem-PAL: Towards Memory-based Personalized Dialogue Assistants for Long-term User-Agent Interaction (AAAI 2026 · similarity 0.79)
- CORECODE: A Common Sense Annotated Dialogue Dataset with Benchmark Tasks for Chinese Large Language Models (AAAI 2024 · similarity 0.79)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
