Research map / Label-efficient and robust learning / Recognition and long-tail
Disentangled CVAEs with Contrastive Learning for Explainable Recommendation
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 disentangled CVAE with CSS self-regularization rather than PETER | explanation generation · Yelp | BLEU-1: 16.93(±0.5) -> 20.11(±0.8) | Helped |
| Use disentangled CVAE with CSS self-regularization rather than PETER | explanation generation · Amazon | BLEU-1: 16.94(±0.4) -> 18.51(±0.9) | Helped |
| Use disentangled CVAE with CSS self-regularization rather than PETER | explanation generation · TripAdvisor | BLEU-1: 19.24(±0.8) -> 22.76(±1.4) | Helped |
| Add Factor Disentanglement (ablation) | disentangled CVAE · Yelp | BLEU-1: 19.37 (↓3.68%) -> 20.11 | Helped |
| Add Factor Disentanglement (ablation) | disentangled CVAE · Yelp | BLEU-4: 1.41 (↓6.00%) -> 1.50 | Helped |
| Add CSS self-regularization loss (ablation) | disentangled CVAE · Yelp | BLEU-1: 18.25 (↓9.25%) -> 20.11 | Helped |
| Add CSS self-regularization loss (ablation) | disentangled CVAE · Yelp | BLEU-4: 1.27 (↓15.33%) -> 1.50 | Helped |
| Add Factor Disentanglement and CSS self-regularization (ablation) | disentangled CVAE · Yelp | METEOR: 5.92 (↓14.68%) -> 6.95 | Helped |
| Replace CSS with hierarchical disentanglement (ablation) | disentangled CVAE · Yelp | BLEU-1: 3.5% | Hurt |
| Replace CSS with SimCSE or vanilla contrastive learning loss (ablation) | disentangled CVAE · Yelp | explanation generation performance | Hurt |
Closest papers by meaning
- DR-VAE: Debiased and Representation-enhanced Variational Autoencoder for Collaborative Recommendation (AAAI 2025 · similarity 0.86)
- Graph Disentangled Contrastive Learning with Personalized Transfer for Cross-Domain Recommendation (AAAI 2024 · similarity 0.85)
- Contrastive Representation for Interactive Recommendation (AAAI 2025 · similarity 0.84)
- DisCo: Graph-Based Disentangled Contrastive Learning for Cold-Start Cross-Domain Recommendation (AAAI 2025 · similarity 0.84)
- Coherency Improved Explainable Recommendation via Large Language Model (AAAI 2025 · similarity 0.84)
- Disentangled Contrastive Bundle Recommendation with Conditional Diffusion (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.