HEAT: A Highly Efficient and Affordable Training System for Collaborative Filtering Based Recommendation on CPUs.
Chengming ZhangShaden SmithBaixi SunJiannan TianJonathan SoiferXiaodong YuShuaiwen Leon SongYuxiong HeDingwen TaoPublished in: ICS (2023)
Keyphrases
- highly efficient
- collaborative filtering
- recommender systems
- recommendation systems
- matrix factorization
- personalized recommendation
- user preferences
- data sparsity
- cold start problem
- low cost
- recommendation quality
- recommendation algorithms
- parallel computing
- online dating
- low latency
- product recommendation
- low complexity
- user profiles
- gray code
- collaborative filtering recommendation
- computational complexity
- user ratings
- discriminative learning
- information filtering
- real time
- tensor factorization
- latent factor models
- training set
- item recommendation
- hybrid recommendation