Research map / Vision-language and multimodal / Cross-modal learning
MusicRec: Multi-modal Semantic-Enhanced Identifier with Collaborative Signals for Generative Recommendation
Yuqiu Zhao, Lei Shi, Yan Zhong, Feifei Kou, Pengfei Zhang, Jiwei Zhang, Mingying Xu, Yanchao Liu
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 |
|---|---|---|---|
| Add image-based identifiers (ablation) | MusicRec · Instruments | HR@10: 0.1172 -> 0.1472 | Improved |
| Add text-based identifiers (ablation) | MusicRec · Games | HR@10: 0.0847 -> 0.1353 | Improved |
| Add collaborative alignment loss (ablation) | MusicRec · Instruments | HR@10: 0.1216 -> 0.1472 | Improved |
| Add collaborative relational preservation loss (ablation) | MusicRec · Games | HR@10: 0.1288 -> 0.1353 | Improved |
| Add multi-modal next item generation task (ablation) | MusicRec · Arts | HR@10: 0.1435 -> 0.1451 | Improved |
| Add multi-modal item alignment tasks (ablation) | MusicRec · Games | HR@10: 0.1349 -> 0.1353 | No clear effect |
| Add multi-modal item alignment tasks (ablation) | MusicRec · Instruments | HR@10: 0.1423 -> 0.1472 | Improved |
| Use MusicRec instead of best baseline | MusicRec · Instruments | HR@10: 0.1361 -> 0.1472 | Improved |
| Use MusicRec instead of MQL4GRec | MusicRec · Games | HR@10: 0.0932 -> 0.1353 | Improved |
| Increase codebook size beyond 256 (ablation) | MusicRec · Arts | recommendation performance | Hurt |
| Use shared:specific codebook ratio 3:1 instead of 1:1 (ablation) | MusicRec · Arts | recommendation performance | Hurt |
| Raise alpha beyond 1e-2 (ablation) | MusicRec · Instruments, Arts, Games | recommendation performance | Hurt |
Closest papers by meaning
- Rethinking Semantic ID Construction for Generative Recommendation: SimHash with Parallel Decoding and Semantic Alignment (NeurIPS 2026 · similarity 0.88)
- Towards Unified Multi-Modal Personalization: Large Vision-Language Models for Generative Recommendation and Beyond (ICLR 2024 · similarity 0.86)
- Multimodal Quantitative Language for Generative Recommendation (ICLR 2025 · similarity 0.85)
- Language Models as Semantic Indexers (ICML 2024 · similarity 0.84)
- MTSTRec: Multimodal Time-Aligned Shared Token Recommender (ICML 2025 · similarity 0.84)
- Semantic Convergence: Harmonizing Recommender Systems via Two-Stage Alignment and Behavioral Semantic Tokenization (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.
