Research map / Probabilistic and scientific modeling / Variational and generative sampling
Towards Diverse, Relevant and Coherent Open-Domain Dialogue Generation via Hybrid Latent Variables
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 |
|---|---|---|---|
| Add hybrid continuous and discrete latent variables | Transformer dialogue generation · DailyDialog | BLEU-4: 0.1412 -> 0.1847 | Helped |
| Add hybrid continuous and discrete latent variables | Transformer dialogue generation · OpenSubtitles | BLEU-4: 0.1112 -> 0.2002 | Helped |
| Add HLV during fine-tuning | PLATO · DailyDialog | BLEU-4: 0.1612 -> 0.1966 | Helped |
| Add HLV during fine-tuning | BART-base · OpenSubtitles | BLEU-4: 0.0565 -> 0.0901 | Helped |
| Add relaxed KL divergence (ablation) | CHVT · DailyDialog | Distinct-3: 0.3074 -> 0.3918 | Helped |
| Add relaxed KL divergence (ablation) | CHVT · OpenSubtitles | BLEU-4: 0.2002 -> 0.2164 | Helped |
| Add relaxed KL divergence (ablation) | CHVT · OpenSubtitles | Distinct-3: 0.1906 -> 0.1516 | Hurt |
| Add continuous latent variables (ablation) | CHVT · DailyDialog test | Distinct-3: 0.1105 -> 0.3074 | Helped |
| Add continuous latent variables (ablation) | CHVT · DailyDialog test | BLEU-4: 0.1711 -> 0.1847 | Helped |
| Add discrete latent variables (ablation) | CHVT · DailyDialog test | Distinct-3: 0.0277 -> 0.3074 | Helped |
| Add discrete latent variables (ablation) | CHVT · DailyDialog test | BLEU-4: 0.1033 -> 0.1847 | Helped |
Closest papers by meaning
- A Hierarchical Latent Variable Encoder-Decoder Model for Generating Dialogues (AAAI 2017 · similarity 0.89)
- Improving Variational Encoder-Decoders in Dialogue Generation (AAAI 2018 · similarity 0.89)
- DialogWAE: Multimodal Response Generation with Conditional Wasserstein Auto-Encoder (ICLR 2019 · similarity 0.89)
- Diverse and Faithful Knowledge-Grounded Dialogue Generation via Sequential Posterior Inference (ICML 2023 · similarity 0.89)
- Improving Open-Domain Dialogue Response Generation with Multi-Source Multilingual Commonsense Knowledge (AAAI 2024 · similarity 0.87)
- SD-Eval: A Benchmark Dataset for Spoken Dialogue Understanding Beyond Words (NeurIPS 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.