resources of multi-arm bandit
- (F1) only one machine is operated at each time instant. The evolution of the machine that is being operated is uncontrolled; that is, the processor chooses which machine to operate but not how to operate it;
- (F2) machines that are not operated remain frozen;
- (F3) machines are independent;
- (F4) frozen machines contribute no reward.
- Epsilon-Greedy
- UCB
- Contextual Bandits
- LinUCB
- CoLin
- hLinUCB
- FactorUCB
- Thompson Sampling (Bayesian)
- Bernoulli, Binomial <=> Beta Distributions
- Reinforcement Learning: An Introduction
- Multi-armed Bandit Allocation Indices
- Bandit Algorithms for Website Optimization
- Multi-Armed Bandit Problems (in Foundations and Applications of Sensor Management)
- Latent Contextual Bandits and Their Application to Personalized Recommendations for New Users
- A Survey on Contextual Multi-armed Bandits
- Contextual Bandits in A Collaborative Environment(SIGIR'2016)
- Learning Hidden Features for Contextual Bandits. (CIKM 2016)
- Factorization Bandits for Interactive Recommendation.(AAAI 2017)
- Returning is Believing: Optimizing Long-term User Engagement in Recommender Systems.(CIKM 2017)
- [Portfolio Choices with Orthogonal Bandit Learning-IJCAI 2015] (http://yugangjiang.info/publication/ijcai15-OBL.pdf)
- When to Run Bandit Tests Instead of A/B/n Tests
- Bandit theory, part I
- Bandit theory, part II
- Bandits for Recommendation Systems
- Recommendations with Thompson Sampling
- Personalization with Contextual Bandits
- Bayesian Bandits - optimizing click throughs with statistics
- Mulit-Armed Bandits
- Bayesian Bandits
- Python Multi-armed Bandits (and Beer!)
- Boston Bayesians Meetup 2016 - Bayesian Bandits From Scratch
- ODSC East 2016 - Bayesian Bandits
- NYC ML Meetup 2010 - Learning for Contextual Bandits
- David S. Leslie-Lancaster University
- B. Van Roy-Stanford University
- Rémi Munos-Deepmind
- Csaba Szepesvari-Deepmind/Alberta University
- Emma Brunskill-Stanford University
- Sebastien Bubeck-Senior Researcher, Theory Group, Microsoft Research
- Nicolò Cesa-Bianchi-Professor of Computer Science, Università degli Studi di Milano
- Vianney Perchet-Professeur de Mathématiques Appliquées au CMLA, ENS Paris-Saclay