Research map / Label-efficient and robust learning / Recognition and long-tail
Stochastically Robust Personalized Ranking for LSH Recommendation Retrieval
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 |
|---|---|---|---|
| learn user/item vectors that are robust to LSH stochasticity (SRPR) instead of plain ranking-based MF | SRPR-SRP vs IBPR · MovieLens 20M Hamming ranking b=8 | Absolute LSH nDCG@10: 0.679 -> 0.697 | Helped |
| learn LSH-robust vectors (SRPR) instead of post-hoc QNF transformation of BPR | BPR-QNF · Netflix Hamming ranking b=128 | Absolute LSH nDCG@10: 0.514 -> 0.752 | Helped |
| learn LSH-compatible real-valued vectors (SRPR-SRP) instead of learning binary codes directly (DPR) | DPR · MovieLens 20M Hamming ranking b=8 | nDCG@10: 0.652 -> 0.697 (+6.9%) | Helped |
| learn LSH-compatible real-valued vectors (SRPR-SRP) instead of learning binary codes directly (DPR) | DPR · Netflix Hamming ranking b=128 | nDCG@10: 0.732 -> 0.752 (+2.7%) | Helped |
| incorporate LSH stochasticity into learning | SRPR vs IBPR/IPMF · MovieLens/Netflix brute-force (no LSH) | nDCG@10: 0.753 vs 0.752 | No clear effect |
| increase code length b | hash-table retrieval · MovieLens 20M / Netflix | nRecall@10 and speedup | Mixed |
Closest papers by meaning
- MONGOOSE: A Learnable LSH Framework for Efficient Neural Network Training (ICLR 2021 · similarity 0.87)
- Locality-Sensitive Hashing for f-Divergences: Mutual Information Loss and Beyond (NeurIPS 2019 · similarity 0.87)
- Optimizing Recall in Deep Graph Hashing Framework for Item Retrieval (Student Abstract) (AAAI 2024 · similarity 0.87)
- Large-Scale Distributed Learning via Private On-Device LSH (NeurIPS 2023 · similarity 0.86)
- Learning to Hash Robustly, Guaranteed (ICML 2022 · similarity 0.86)
- From Pairwise to Ranking: Climbing the Ladder to Ideal Collaborative Filtering with Pseudo-Ranking (AAAI 2025 · 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.