Research map / Point clouds / Shape and surface completion
Towards a Neural Conversation Model With Diversity Net Using Determinantal Point Processes
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 diversity net to DPP-D decoder (ablation) | DPP-D · Baidu Tieba test | BLEU-1 (top-1): 3.157 -> 8.568 | Helped |
| add diversity net to DPP-D decoder (ablation) | DPP-D · Baidu Tieba test | human quality score: 0.927 -> 1.440 | Helped |
| add diversity net to DPP-D decoder (ablation) | DPP-D · Baidu Tieba test | distinct-4: 0.572 -> 0.809 | Helped |
| add diversity net to DPP-D decoder (ablation) | DPP-D · Baidu Tieba test | universal reply percentage: 7.333% -> 5.333% | Helped |
| apply DPP at every decoding step instead of re-ranking after beam search (ablation) | DPP-R · Baidu Tieba test | BLEU-1 (top-1): 2.698 -> 3.157 | Helped |
| apply DPP at every decoding step instead of re-ranking after beam search (ablation) | DPP-R · Baidu Tieba test | distinct-1: 0.026 -> 0.023 | Mixed |
| DPP re-ranking of beam search candidates | SBS · Baidu Tieba test | BLEU-1 (top-1): 1.053 -> 2.698 | Helped |
| DPP re-ranking of beam search candidates | SBS · Baidu Tieba test | distinct-4: 0.076 -> 0.358 | Helped |
| diverse beam search instead of standard beam search | SBS · Baidu Tieba test | BLEU-1 (top-1): 1.053 -> 0.363 | Hurt |
| diverse beam search instead of standard beam search | SBS · Baidu Tieba test | distinct-4: 0.076 -> 0.547 | Helped |
| full DPP-D-DivNet instead of standard beam search | SBS · Baidu Tieba test | universal reply percentage: 55.33% -> 5.333% | Helped |
Closest papers by meaning
- A Knowledge-Grounded Neural Conversation Model (AAAI 2018 · similarity 0.88)
- A Statistical Framework for Ranking LLM-based Chatbots (ICLR 2025 · similarity 0.85)
- SimpleStrat: Diversifying Language Model Generation with Stratification (NeurIPS 2025 · similarity 0.83)
- WildBench: Benchmarking LLMs with Challenging Tasks from Real Users in the Wild (ICLR 2025 · similarity 0.83)
- WildChat-50M: A Deep Dive Into the Role of Synthetic Data in Post-Training (ICML 2025 · similarity 0.83)
- MixEval: Deriving Wisdom of the Crowd from LLM Benchmark Mixtures (NeurIPS 2024 · similarity 0.83)
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.