This website collects recent works and datasets on learning to hash for recommendation and their codes. We hope this website could help you do search on this topic.
If you find our project is useful to your research or work, please give us a star ⭐ on GitHub for the latest update and cite our paper. Thank you!
Authors: Fangyuan Luo1, Honglei Zhang2, Tong Li1, Jun Wu2
Affliation: 1Beijing University of Technology, 2Beijing Jiaotong University
- A. Singh and S. Gupta, “Learning to hash: a comprehensive survey of deep learning-based hashing methods,” Knowledge and Information Systems, vol. 64, no. 10, pp. 2565–2597, 2022.
- J. Wang, W. Liu, S. Kumar, and S. Chang, “Learning to hash for indexing big data - A survey,” Proceedings of the IEEE, vol. 104, no. 1, pp. 34–57, 2016.
- J. Wang, T. Zhang, J. Song, N. Sebe, and H. T. Shen, “A survey on learning to hash,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 40, no. 4, pp. 769–790, 2018.
- X.Luo,H.Wang,D.Wu,C.Chen,M.Deng,J.Huang,andX.Hua,“A survey on deep hashing methods,” ACM Transactions on Knowledge Discovery from Data, vol. 17, no. 1, pp. 15:1–15:50, 2023.
- Z. Li, H. Li, and L. Meng, “Model compression for deep neural networks: A survey,” Computers, vol. 12, no. 3, p. 60, 2023.
| Paper | Model | Venue | Task | Learning Objective | Optimization Strategy | Code | |
|---|---|---|---|---|---|---|---|
| Learning binary codes for collaborative filtering | BCCF | KDD'12 | User-Item CF | Pointwise, Pairwise | Two-Stage | Code | |
| Collaborative Hashing | CH | CVPR'14 | User-Item CF | Pointwise | Two-Stage | Code | |
| Preference preserving hashing for efficient recommendation | PPH | SIGIR'14 | User-Item CF | Pointwise | Two-Stage | Code | |
| Constraint Free Preference Preserving Hashing for Fast Recommendation | NQ | IEEE GLOBECOM'16 | User-Item CF | Pointwise | Two-Stage | Code | |
| Discrete Collaborative Filtering | DCF | SIGIR'14 | User-Item CF | Pointwise | One-Stage | Code | |
| Discrete Content-aware Matrix Factorization | DCMF | KDD'17 | Cold-Start | Pointwise | One-Stage | Code | |
| Discrete Personalized Ranking for Fast Collaborative Filtering from Implicit Feedback | DPR | AAAI'17 | User-Item CF | Pairwise | One-Stage | Code | |
| Discrete Factorization Machines for Fast Feature-based Recommendation | DFM | IJCAI'18 | Cold-Start | Pointwise | One-Stage | Code | |
| Learning Discrete Hashing Towards Efficient Fashion Recommendation | DSFCH | Data Science and Engineering'18 | Outfit Recommendation | Pointwise | One-Stage | Code | |
| Discrete Ranking-based Matrix Factorization with Self-Paced Learning | DRMF | KDD'18 | User-Item CF | Pointwise | One-Stage | Code | |
| Discrete Deep Learning for Fast Content-Aware Recommendation | DDL | WSDM'18 | Cold-Start | Pointwise | One-Stage | Code | |
| Discrete Trust-aware Matrix Factorization for Fast Recommendation | DTMF | IJCAI'19 | Social Recommendation | Pointwise | One-Stage | Code | |
| Candidate Generation with Binary Codes for Large-Scale Top-N Recommendation | CIGAR | CIKM'19 | User-Item CF | Pairwise | Two-Stage | Code | |
| Compositional Coding for Collaborative Filtering | CCCF | SIGIR'19 | User-Item CF | Pointwise | One-Stage | Code | |
| Discrete Social Recommendation | DSR | AAAI'19 | Social Recommendation | Pointwise | One-Stage | Code | |
| Learning Binary Code for Personalized Fashion Recommendation | FHN | CVPR'19 | Outfit Recommendation | Pairwise | Two-Stage | Code | |
| Binarized Collaborative Filtering with Distilling Graph Convolutional Networks | DGCN-BinCF | IJCAI'19 | User-Item CF | Pairwise | Two-Stage | Code | |
| Adversarial Binary Collaborative Filtering for Implicit Feedback | ABinCF | AAAI'19 | User-Item CF | Pointwise | Two-Stage | Code | |
| Content-aware Neural Hashing for Cold-start Recommendation | NeuHash-CF | SIGIR'20 | Cold-Start | Pointwise | Two-Stage | Code | |
| Learning to Hash with Graph Neural Networks for Recommender Systems | HashGNN | WWW'20 | User-Item CF | Heterogeneous | Two-Stage | Code | |
| Semi-discrete Matrix Factorization | SDMF | IEEE Intelligent Systems'20 | User-Item CF | Pointwise | One-Stage | Code | |
| Multi-Feature Discrete Collaborative Filtering for Fast Cold-start Recommendation | MFDCF | AAAI'20 | Cold-Start | Pointwise | One-Stage | Code | |
| Collaborative Generative Hashing for Marketing and Fast Cold-Start Recommendation | CGH | IEEE Intelligent Systems'20 | Cold-Start | Pointwise | Two-Stage | Code | |
| Deep Pairwise Hashing for Cold-start Recommendation | DPH | IEEE TKDE'22 | Cold-Start | Pairwise | One-Stage | Code | |
| Projected Hamming Dissimilarity for Bit-Level Importance Coding in Collaborative Filtering | VH_{PHD} | WWW'21 | User-Item CF | Pointwise | Two-Stage | Code | |
| Discrete Matrix Factorization and Extension for Fast Item Recommendation | DMF | IEEE TKDE'21 | Cold-Start | Pointwise | One-Stage | Code | |
| Discrete Listwise Collaborative Filtering for Fast Recommendation | DLCF | SDM'21 | User-Item CF | Listwise | Proximal One-Stage | Code | |
| Semi-Discrete Social Recommendation (Student Abstract) | SDSR | AAAI'21 | Social Recommendation | Pointwise | One-Stage | Code | |
| Learning Binarized Graph Representations with Multi-faceted Quantization Reinforcement for Top-K Recommendation | BiGeaR | KDD'22 | User-Item CF | Pairwise | Two-Stage | Code | |
| Bi-directional Heterogeneous Graph Hashing towards Efficient Outfit Recommendation | BIHGH | ACM MM'22 | Outfit Recommendation | Pairwise | Two-Stage | Code | |
| Discrete Listwise Personalized Ranking for Fast Top-N Recommendation with Implicit Feedback | DLPR | IJCAI'22 | User-Item CF | Listwise | One-Stage | Code | |
| HCFRec Hash Collaborative Filtering via Normalized Flow with Structural Consensus for Efficient Recommendation | HCFRec | IJCAI'22 | User-Item CF | Pointwise | Two-Stage | Code | |
| Explainable discrete Collaborative Filtering | EDCF | IEEE TKDE'22 | Explainable Recommendation | Pointwise | Two-Stage | Code | |
| Discrete Limited Attentional Collaborative Filtering for Fast Social Recommendation | DLACF | EAAI'23 | Social Recommendation | Pointwise | Two-Stage | Code | |
| Personalized Fashion Recommendation With Discrete Content-Based Tensor Factorization | FHN+ | IEEE TMM'23 | Outfit Recommendation | Pairwise | Two-Stage | Code | |
| Hashing-based semantic relevance attributed knowledge graph embedding enhancement for deep probabilistic recommendation | H-SAGE | Applied Intelligence'23 | KG-based Recommendation | Pointwise | Two-Stage | Code | |
| Bipartite Graph Convolutional Hashing for Effective and Efficient Top-N Search in Hamming Space | BGCH | WWW'23 | User-Item CF | Heterogeneous | Two-Stage | Code | |
| Multi-Modal Discrete Collaborative Filtering for Efficient Cold-Start Recommendation | MDCF | IEEE TKDE'23 | Cold-Start | Pointwise | One-Stage | Code | |
| LightFR: Lightweight Federated Recommendation with Privacy-preserving Matrix Factorization | LightFR | ACM TOIS'23 | Federated Recommendation | Pointwise | One-Stage | Code | |
| Discrete Listwise Content-aware Recommendation | DLFM | ACM TKDD'24 | Cold-Start | Listwise | Proximal One-Stage | Code | |
| Discrete Federated Multi-behavior Recommendation for Privacy-Preserving Heterogeneous One-Class Collaborative Filtering | DFMR | ACM TOIS'24 | Federated Recommendation | Pointwise | One-Stage | Code | |
| Temporal Social Graph Network Hashing for Efficient Recommendation | TSGNH | IEEE TKDE'24 | Social Recommendation | Pointwise | Two-Stage | Code | |
| Towards Effective Top-N Hamming Search via Bipartite Graph Contrastive Hashing | BGCH+ | IEEE TKDE'24 | User-Item CF | Pairwise | Two-Stage | Code |
We contain the implementation of evaluation metrics for recommender systems. Suppose that there are two users. The label and prediction list are [[1,1,1,1,0,0,0,0,1,1,1,1], [1,1,0,0,1]] and [[1,1,0,1,0,0,0,0,0,0,1,0], [0,1,1,0,0]] respectively, where each sublist in label or prediction corresponds to one user's label list or prediction list. And K denotes the cut-off of the list.
from RecMetrics import Metric
metric = Metric(label,pred, K)
hitratio = metric.hit_ratio()
recall = metric.recall()
accuracy = metric.accuracy()
mae = metric.mae()
rmse = metric.rmse()
ndcg = metric.ndcg()
map = metric.map()
auc = metric.auc()
mrr = metric.mrr()We collect some datasets (Download) which are often used in the research of HashRec.
| Datasets | #Users | #Items | #Interactions | Density |
|---|---|---|---|---|
| MovieLens 1M | 6,040 | 3,952 | 1,000,209 | 4.19% |
| Movielens 10M | 71,567 | 10,681 | 10,000,054 | 1.31% |
| EachMovie | 1,648 | 74,424 | 2,811,717 | 2.83% |
| Netflix | 480,189 | 17,770 | 100,480,507 | 1.18% |
| Yelp | 13,679 | 12,922 | 640,143 | 0.36% |
| Amazon Book | 35,151 | 33,195 | 1,732,060 | 0.15% |
| Gowalla | 29,858 | 40,981 | 1,027,370 | 0.08% |
| 55,186 | 9,916 | 1,463,556 | 0.27% | |
| BookCrossing | 278,858 | 271,379 | 1,149,780 | 0.002% |
🔥🔥 We will keep updating this list, and if you find any missing related work or have any suggestions, please feel free to contact us (luofangyuan@bjut.edu.cn).
@article{Luo2024HashRec,
title = {Learning to Hash for Recommendation: A Survey},
author = {Fangyuan Luo, Honglei Zhang, Tong Li and Jun Wu},
year = {2024},
journal = {arXiv preprint arXiv: 2412.03875}
}