Research map / Large language models / Language, safety and interpretability
MOSS: End-to-End Dialog System Framework with Modular Supervision
Weixin Liang, Youzhi Tian, Chengcai Chen, Zhou Yu
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 |
|---|---|---|---|
| add supervision from all four dialog modules | MOSS · CamRest676 test | Mat: 0.927 -> 0.951 | Improved |
| add supervision from dialog policy learning (ablation) | MOSS · CamRest676 test | BLEU: 0.251 -> 0.259 | Improved |
| add supervision from natural language understanding (ablation) | MOSS · CamRest676 test | Mat: 0.932 -> 0.951 | Improved |
| add 40% raw dialogs to 60% fully annotated training data (ablation) | MOSS-all · CamRest676 test | BLEU: 0.202 -> 0.221 | Improved |
| reduce training data to 60% (ablation) | MOSS-all · CamRest676 test | BLEU: 0.259 -> 0.202 | Hurt |
| reduce training data to 60% (ablation) | MOSS-all · CamRest676 test | Mat: 0.951 -> 0.947 | No clear effect |
| add supervision from dialog policy learning (ablation) | MOSS · LaptopNetwork test | Succ.acc | Improved |
| add supervision from natural language understanding (ablation) | MOSS · LaptopNetwork test | Succ.acc: 0.50 -> 0.61 | Improved |
| use modular supervision from all four modules rather than TSCP supervision | MOSS-all · LaptopNetwork test | Succ.acc: 0.18 -> 0.61 | Improved |
| reduce training data to 40% (ablation) | MOSS-all · LaptopNetwork test | Succ.acc | Improved |
Closest papers by meaning
- TA&AT: Enhancing Task-Oriented Dialog with Turn-Level Auxiliary Tasks and Action-Tree Based Scheduled Sampling (AAAI 2024 · similarity 0.86)
- Dialog-based Language Learning (NeurIPS 2016 · similarity 0.86)
- An Efficient Self-Learning Framework For Interactive Spoken Dialog Systems (ICML 2024 · similarity 0.84)
- Winnie: Task-Oriented Dialog System with Structure-Aware Contrastive Learning and Enhanced Policy Planning (AAAI 2024 · similarity 0.83)
- Zhongjing: Enhancing the Chinese Medical Capabilities of Large Language Model through Expert Feedback and Real-World Multi-Turn Dialogue (AAAI 2024 · similarity 0.82)
- StateTree: Enhancing Long-term Dialogue Reasoning via Reinforcement Learning (NeurIPS 2026 · similarity 0.82)
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.
