Research map / Large language models / Language, safety and interpretability
A Simple Language Model for Task-Oriented Dialogue
Ehsan Hosseini-Asl, Bryan McCann, Chien-Sheng Wu, Semih Yavuz, Richard Socher
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 |
|---|---|---|---|
| single causal LM for DST, action and response (SimpleTOD) | GPT-2 · MultiWOZ 2.1 DST | joint accuracy: 55.3 -> 55.72 | Improved |
| SimpleTOD vs DAMD with oracle DB | GPT-2 · MultiWOZ 2.0 end-to-end | combined score: 85 -> 87.66 | Improved |
| use oracle DB search results as input (ablation) | SimpleTOD · MultiWOZ 2.1 end-to-end | combined score: 92.98 -> 87.36 | Hurt |
| use dynamic DB search results (ablation) | SimpleTOD · MultiWOZ 2.1 end-to-end | combined score: 92.98 -> 90.24 | Hurt |
| add end-of-segment and user/system tokens (ablation) | SimpleTOD · MultiWOZ 2.1 | combined score: 26.73 -> 92.98 | Improved |
| add user/system tokens (ablation) | SimpleTOD with end tokens · MultiWOZ 2.1 | joint accuracy: 21.5 -> 55.76 | Improved |
| add end-of-segment tokens (ablation) | SimpleTOD with user/system tokens · MultiWOZ 2.1 | combined score: 66.87 -> 92.98 | Improved |
| initialize from pretrained DistilGPT2 (ablation) | 6-layer SimpleTOD · MultiWOZ 2.1 | joint accuracy: 16.45 -> 54.54 | Improved |
| initialize from pretrained GPT2 (ablation) | 12-layer SimpleTOD · MultiWOZ 2.1 | combined score: 59.65 -> 90.75 | Improved |
| train three separate LMs per sub-task instead of one (ablation) | SimpleTOD · MultiWOZ | combined score | No clear effect |
| nucleus sampling instead of greedy decoding (ablation) | SimpleTOD · MultiWOZ | overall performance | Hurt |
| 12 layers GPT2 vs 6 layers DistilGPT2 (ablation) | SimpleTOD · MultiWOZ 2.1 | combined score: 92.98 -> 90.75 | Mixed |
Closest papers by meaning
- SpokenWOZ: A Large-Scale Speech-Text Benchmark for Spoken Task-Oriented Dialogue Agents (NeurIPS 2023 · similarity 0.88)
- TA&AT: Enhancing Task-Oriented Dialog with Turn-Level Auxiliary Tasks and Action-Tree Based Scheduled Sampling (AAAI 2024 · similarity 0.87)
- From Retrieval to Generation: A Simple and Unified Generative Model for End-to-End Task-Oriented Dialogue (AAAI 2024 · similarity 0.86)
- StateTree: Enhancing Long-term Dialogue Reasoning via Reinforcement Learning (NeurIPS 2026 · similarity 0.85)
- ToolDial: Multi-turn Dialogue Generation Method for Tool-Augmented Language Models (ICLR 2025 · similarity 0.85)
- TaskLAMA: Probing the Complex Task Understanding of Language Models (AAAI 2024 · similarity 0.85)
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.
