Research map / Architectures and efficiency / RNNs and normalization
Hyperbolic RQ-VAE enhanced Generative Recommendation with Differential-Length Codebook Strategy
Aoran Zhang, Yu-Bin Yang, Yonghong Yu
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 hyperbolic RQ-VAE (to codebook [256,256,256]) (ablation) | HG-Rec · Beauty | NDCG@10: 0.0418 -> 0.0457 | Improved |
| add hyperbolic RQ-VAE (to codebook [256,256,256]) (ablation) | HG-Rec · Yelp | NDCG@10: 0.0203 -> 0.0222 | Improved |
| add differential-length codebook [64,128,256] (ablation) | HG-Rec · Beauty | NDCG@10: 0.0424 -> 0.0473 | Improved |
| add differential-length codebook to TIGER (ablation) | TIGER · Beauty | NDCG@10: 0.0408 -> 0.0424 | Improved |
| add differential-length codebook to TIGER (ablation) | TIGER · Yelp | NDCG@10: 0.0203 -> 0.0209 | No clear effect |
| reduce codebook size from [256,256,256] to [64,64,64] (ablation) | TIGER · Beauty | NDCG@10: 0.0418 -> 0.0408 | Hurt |
| add HG-Rec over ActionPiece | ActionPiece · Yelp | NDCG@10: 0.0219 -> 0.0233 | Improved |
| add HG-Rec over ActionPiece | ActionPiece · Beauty | NDCG@5: 0.0357 -> 0.0377 | Improved |
| replace vanilla RQ-VAE with hyperbolic RQ-VAE | RQ-VAE tokenizer · Beauty (collision rate) | collision rate | Improved |
| replace vanilla RQ-VAE with hyperbolic RQ-VAE | RQ-VAE tokenizer · Beauty (codebook usage) | codebook usage | Improved |
| replace vanilla RQ-VAE with hyperbolic RQ-VAE | RQ-VAE tokenizer · Beauty (codebook [256,256,256]) | total training time: 3724s -> 103s | Improved |
| replace vanilla RQ-VAE with hyperbolic RQ-VAE | RQ-VAE tokenizer · Yelp (codebook [256,256,256]) | total training time: 5445s -> 144s | Improved |
Closest papers by meaning
- Efficient Inference for Large Language Model-based Generative Recommendation (ICLR 2025 · similarity 0.87)
- HELM: Hyperbolic Large Language Models via Mixture-of-Curvature Experts (NeurIPS 2025 · similarity 0.85)
- Actions Speak Louder than Words: Trillion-Parameter Sequential Transducers for Generative Recommendations (ICML 2024 · similarity 0.85)
- Hgformer: Hyperbolic Graph Transformer for Collaborative Filtering (ICML 2025 · similarity 0.85)
- Hyperbolic Fine-Tuning for Large Language Models (NeurIPS 2025 · similarity 0.84)
- Multimodal Quantitative Language for Generative Recommendation (ICLR 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.
