Research map / Label-efficient and robust learning / Recognition and long-tail
HVAE: Hyperbolic Variational Autoencoder For Flexible Knowledge Transfer Across Multiple Domains
Xiaolei Liu, Binfeng Wang, Kaixin Gao, Shaoshuai Li
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 hyperbolic space instead of Euclidean embeddings (ablation) | HVAE · MYbank and Amazon CDR | H@10 / N@10 | Improved |
| add learnable domain-specific curvature instead of fixed curvature (ablation) | HVAE · MYbank and Amazon CDR | H@10 / N@10 | Improved |
| use variational autoencoder instead of deterministic autoencoder (ablation) | HVAE · MYbank and Amazon CDR | H@10 / N@10 | Improved |
| add cycle pseudo-labeling self-training (ablation) | HVAE · MYbank and Amazon CDR | H@10 / N@10 | Improved |
| add hyperbolic Wasserstein distance disentanglement loss and domain classification loss (ablation) | HVAE · MYbank and Amazon CDR | H@10 / N@10 | Improved |
| lower user overlap ratio from 70% to 1% | HVAE vs strongest baseline · MYbank, Amazon | H@10 / N@10 gain over baseline: 3.5%-10.46% (H@10), 4.31%-15.16% (N@10) at 1% overlap on MYbank | Improved |
| deploy HVAE versus BPR, LightGCN and other baselines online | HVAE · Ant Group online A/B test | CTR: 7.453% average improvement | Improved |
| use standard Adam instead of Riemannian Adam | HVAE · optimization | stability and efficiency | Improved |
Closest papers by meaning
- Hyperbolic RQ-VAE enhanced Generative Recommendation with Differential-Length Codebook Strategy (ICML 2026 · similarity 0.87)
- HELM: Hyperbolic Large Language Models via Mixture-of-Curvature Experts (NeurIPS 2025 · similarity 0.86)
- Hgformer: Hyperbolic Graph Transformer for Collaborative Filtering (ICML 2025 · similarity 0.85)
- Graph Geometry-Preserving Autoencoders (ICML 2024 · similarity 0.85)
- Low-distortion and GPU-compatible Tree Embeddings in Hyperbolic Space (ICML 2025 · similarity 0.84)
- Hyperbolic Fine-Tuning for Large Language Models (NeurIPS 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.
