Research map / Large language models / Language, safety and interpretability
Topic Aware Neural Response Generation
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 |
|---|---|---|---|
| incorporate topic information with joint attention and biased topic-word generation | Seq2Seq response generation · Baidu Tieba human evaluation | +2 responses: 33.1% -> 44.7% | Helped |
| incorporate topic information with joint attention and biased topic-word generation | Seq2Seq response generation · Baidu Tieba test | distinct-1 count: 603 -> 1355 | Helped |
| incorporate topic information with joint attention and biased topic-word generation | Seq2Seq response generation · Baidu Tieba test | distinct-2 count: 1073 -> 2970 | Helped |
| incorporate topic information with joint attention and biased topic-word generation | Seq2Seq response generation · Baidu Tieba test | PPL-T: 133.11 -> 122.82 | Helped |
| add biased generation probability for topic words (ablation) | TA-Seq2Seq · Baidu Tieba human evaluation | +2 responses: 42.3% -> 44.7% | Helped |
| add biased generation probability for topic words (ablation) | TA-Seq2Seq · Baidu Tieba test | distinct-1 count: 894 -> 1355 | Helped |
| add biased generation probability for topic words (ablation) | TA-Seq2Seq · Baidu Tieba test | PPL-T: 119.55 -> 122.82 | Hurt |
| use topic attention instead of a concatenated topic vector (ablation) | topic-aware Seq2Seq · Baidu Tieba human evaluation | +2 responses: 35.9% -> 42.3% | Helped |
| use topic attention instead of a concatenated topic vector (ablation) | topic-aware Seq2Seq · Baidu Tieba test | distinct-1 count: 898 -> 894 | No clear effect |
| add topic information to standard Seq2Seq (ablation) | Seq2Seq response generation · Baidu Tieba human evaluation | +2 responses: 32.3% -> 35.9% | Helped |
Closest papers by meaning
- Neural Response Generation With Dynamic Vocabularies (AAAI 2018 · similarity 0.89)
- Hierarchical Recurrent Attention Network for Response Generation (AAAI 2018 · similarity 0.89)
- A Knowledge-Grounded Neural Conversation Model (AAAI 2018 · similarity 0.88)
- Mechanism-Aware Neural Machine for Dialogue Response Generation (AAAI 2017 · similarity 0.86)
- A Statistical Framework for Ranking LLM-based Chatbots (ICLR 2025 · similarity 0.84)
- From Retrieval to Generation: A Simple and Unified Generative Model for End-to-End Task-Oriented Dialogue (AAAI 2024 · similarity 0.84)
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.