Research map / Graph learning / Knowledge and temporal graphs
Infusing Multi-Source Knowledge with Heterogeneous Graph Neural Network for Emotional Conversation 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 |
|---|---|---|---|
| Use heterogeneous graph encoding instead of ReCoSa with two-source knowledge | MELD response generation · MELD test, two sources | PPL: 108.51 -> 106.04 | Helped |
| Use heterogeneous graph encoding instead of ReCoSa with two-source knowledge | MELD response generation · MELD test, two sources | W-avg.: 30.13 -> 31.58 | Helped |
| Use heterogeneous graph encoding instead of ReCoSa with four-source knowledge | MELD response generation · MELD test, four sources | BLEU: 2.49 -> 2.80 | Helped |
| Use heterogeneous graph encoding instead of ReCoSa with two-source knowledge | DailyDialog response generation · DailyDialog test, two sources | W-avg.: 69.63 -> 71.16 | Helped |
| Add facial-expression nodes (ablation) | HGNN · MELD test, four sources | PPL: 113.94 -> 96.90 | Helped |
| Add facial-expression nodes (ablation) | HGNN · MELD test, four sources | W-avg.: 31.54 -> 34.49 | Helped |
| Add emotion nodes (ablation) | HGNN · MELD test, four sources | PPL: 111.42 -> 96.90 | Helped |
| Add emotion nodes (ablation) | HGNN · MELD test, four sources | W-avg.: 30.18 -> 34.49 | Helped |
| Add speaker nodes (ablation) | HGNN · MELD test, four sources | W-avg.: 32.92 -> 34.49 | Helped |
| Add current-speaker input to the decoder, retaining speaker nodes in HGNN (ablation) | HGNN · MELD test, four sources | W-avg.: 33.24 -> 34.49 | Helped |
Closest papers by meaning
- Emotional Chatting Machine: Emotional Conversation Generation with Internal and External Memory (AAAI 2018 · similarity 0.90)
- Emotion Rendering for Conversational Speech Synthesis with Heterogeneous Graph-Based Context Modeling (AAAI 2024 · similarity 0.88)
- Adaptive Graph Learning for Multimodal Conversational Emotion Detection (AAAI 2024 · similarity 0.87)
- Dynamic Interactive Bimodal Hypergraph Networks for Emotion Recognition in Conversations (AAAI 2025 · similarity 0.85)
- BIG-FUSION: Brain-Inspired Global-Local Context Fusion Framework for Multimodal Emotion Recognition in Conversations (AAAI 2025 · similarity 0.84)
- Dual-View Learning for Conversational Emotion Recognition Through Context and Emotion-Shift Modeling (AAAI 2025 · 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.