Research map / Graph learning / Knowledge and temporal graphs
Entrainment2Vec: Embedding Entrainment for Multi-Party Dialogues
Zahra Rahimi, Diane Litman
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 weakly-supervised learned embedding instead of HAM baseline | entrainment vectors · Conflict prediction (FS) | accuracy: 69.74 -> 74.79 | Improved |
| use weakly-supervised learned embedding instead of HAM baseline | entrainment vectors · Favorable prediction (all features) | accuracy: 53.78 -> 63.86 | Improved |
| kernel initialization instead of uniform initialization (ablation) | Self-supervised embedding (permuted negatives) · Conflict, FS | accuracy: 63.86 -> 73.95 | Improved |
| kernel initialization instead of uniform initialization (ablation) | Self-supervised embedding (permuted negatives) · Favorable, all features | accuracy: 39.49 -> 57.14 | Improved |
| kernel initialization instead of uniform initialization (ablation) | Self-supervised embedding (random negatives) · Favorable, all features | accuracy: 58.82 -> 56.30 | Hurt |
| permuted-conversation negative sampling instead of random negatives (ablation) | Self-supervised kernel embedding · Conflict, FS | accuracy: 66.38 -> 73.95 | Improved |
| permuted negatives instead of random negatives (ablation) | Self-supervised kernel embedding · Favorable, all features | accuracy: 56.30 -> 57.14 | Mixed |
| keep node embeddings fixed (not trainable) in weakly-supervised model (ablation) | WeakS Kernel Permuted · Conflict, FS | accuracy: 63.86 -> 74.79 | Improved |
| keep node embeddings fixed (not trainable) (ablation) | WeakS Kernel Permuted · Favorable, all features | accuracy: 56.30 -> 63.86 | Improved |
| train embedding (weakly-supervised) on top of kernel vectors (ablation) | Kernel · Conflict, FS | accuracy: 69.74 -> 74.79 | Improved |
| kernel vectors without training vs HAM | Kernel · Conflict, all features | accuracy: 68.90 -> 57.14 | Hurt |
| add LASSO feature selection (ablation) | Kernel embedding · Conflict | accuracy: 57.14 -> 69.74 | Improved |
Closest papers by meaning
- Friends-MMC: A Dataset for Multi-modal Multi-party Conversation Understanding (AAAI 2025 · similarity 0.84)
- Modeling Multimodal Social Interactions: New Challenges and Baselines with Densely Aligned Representations (CVPR 2024 · similarity 0.84)
- MultiTalk: Scaling Full-Duplex Speech Models to Long, Multi-Party, Bilingual Conversation (NeurIPS 2026 · similarity 0.83)
- Decoupling Conversational Dynamics in Full-Duplex Spoken Models through Reinforcement Learning (NeurIPS 2026 · similarity 0.83)
- SD-Eval: A Benchmark Dataset for Spoken Dialogue Understanding Beyond Words (NeurIPS 2024 · similarity 0.83)
- The Audio-Visual Conversational Graph: From an Egocentric-Exocentric Perspective (CVPR 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.
