Research map / Probabilistic and scientific modeling / Molecules and proteins
Diverse and Faithful Knowledge-Grounded Dialogue Generation via Sequential Posterior Inference
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| use SPI-uniform instead of KnowledGPT | Knowledge-grounded response generation · WoW Seen | R2: 7.9 -> 8.8 | Helped |
| use SPI-uniform instead of KnowledGPT | Knowledge-grounded response generation · WoW Unseen | R2: 6.7 -> 8.5 | Helped |
| use SPI-uniform instead of KnowledGPT | Knowledge-grounded response generation · WoW Seen | QuestEval RD: 42.2 -> 44.4 | Helped |
| use SPI-uniform instead of KnowledGPT | Knowledge-grounded response generation · WoW Unseen | QuestEval RD: 41.0 -> 44.0 | Helped |
| use SPI-uniform instead of KnowledGPT | Knowledge-grounded response generation · WoW Seen | Fluency: 62.5% -> 88.7% | Helped |
| use SPI-uniform instead of KnowledGPT | Knowledge-grounded response generation · WoW Seen | Faithfulness: 3.33 -> 3.66 | Helped |
| use SPI-uniform instead of CoLV | Knowledge-grounded response generation · Holl-E | Acc: 32.7 -> 38.3 | Helped |
| use SPI-uniform instead of CoLV | Knowledge-grounded response generation · Holl-E | PPL: 34.8 -> 12.6 | Helped |
| increase initializer candidates from top-1 to top-5 (ablation) | SPI-uniform · WoW Seen | B4: 7.3 -> 7.7 | Helped |
| increase initializer candidates from top-5 to top-10 (ablation) | SPI-uniform · WoW Seen | B4: 7.7 -> 7.2 | Hurt |
| increase initializer candidates from top-1 to top-5 (ablation) | SPI-uniform · WoW Seen | Dist-2: 36.6 -> 40.9 | Helped |
| add 5 Langevin posterior-inference steps instead of none (ablation) | SPI-uniform · WoW Seen | B4: 7.4 -> 7.7 | Helped |
Closest papers by meaning
- Sequential Latent Knowledge Selection for Knowledge-Grounded Dialogue (ICLR 2020 · similarity 0.92)
- Zero-Resource Knowledge-Grounded Dialogue Generation (NeurIPS 2020 · similarity 0.90)
- Low-Resource Knowledge-Grounded Dialogue Generation (ICLR 2020 · similarity 0.90)
- From Retrieval to Generation: A Simple and Unified Generative Model for End-to-End Task-Oriented Dialogue (AAAI 2024 · similarity 0.88)
- Improving the Robustness of Knowledge-Grounded Dialogue via Contrastive Learning (AAAI 2024 · similarity 0.88)
- Improving Open-Domain Dialogue Response Generation with Multi-Source Multilingual Commonsense Knowledge (AAAI 2024 · similarity 0.87)
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.