Phase Transitions and Sample Complexity in Bayes-Optimal Matrix Factorization.
Yoshiyuki KabashimaFlorent KrzakalaMarc MézardAyaka SakataLenka ZdeborováPublished in: IEEE Trans. Inf. Theory (2016)
Keyphrases
- matrix factorization
- bayes optimal
- sample complexity
- phase transition
- generalization error
- collaborative filtering
- learning curves
- constraint satisfaction
- upper bound
- vc dimension
- learning curve
- active learning
- pac learning
- recommender systems
- supervised learning
- learning algorithm
- theoretical analysis
- model selection
- cross validation
- sample size
- missing data
- linear classifiers
- training data
- learning problems
- training set
- negative matrix factorization
- lower bound
- np complete
- special case
- random graphs
- concept classes
- training examples
- machine learning
- training samples
- prior knowledge