Aleksandrs Slivkins
Aleksandrs Slivkins is a Microsoft Research researcher and author of Introduction to Multi-Armed Bandits, an introductory textbook-style treatment of bandit algorithms and their connections to machine learning, algorithms, statistics, operations research, and economics.
Key facts
- Type: Researcher / author
- Institution in source: Microsoft Research NYC [src-019]
- Known here for: Introduction to Multi-Armed Bandits, first drafted in 2017, published in 2019, and revised online through April 2024 [src-019]
- Subject area: Multi-Armed Bandits, including Stochastic Bandits, Thompson Sampling, Contextual Bandits, Adversarial Bandits, and Bandits with Knapsacks [src-019]
What it does
In the source, Slivkins frames multi-armed bandits as a teachable family of online decision-making problems rather than a single algorithm. The book decomposes the field into lines of work and presents each as a self-contained technical introduction with pointers to deeper literature [src-019].
The practitioner value of the book is its structure: it separates IID reward settings, Bayesian priors, similarity information, adversarial rewards, contextual signals, and economic constraints. That makes it useful as a map for choosing which bandit model fits a product, marketplace, experimentation, or allocation problem [src-019].
Related
- See also: Multi-Armed Bandits, Exploration-Exploitation Trade-off, Stochastic Bandits, Thompson Sampling, Contextual Bandits, Adversarial Bandits, Bandits with Knapsacks
Source references
- [src-019] Aleksandrs Slivkins — “Introduction to Multi-Armed Bandits” (2019-04-15; revised 2024-04-03)
Keep reading from this thread
From 491 indexed pages and articles.
- Wiki concept Multi-Armed Bandits A framework for algorithms that make repeated decisions under uncertainty, learning from partial feedback while balancing exploration of uncertain actions against Related by multi
- Wiki concept Contextual Bandits Bandit problems where each decision arrives with observable context, and the algorithm learns which actions work best for which contexts from Related by bandits
- Insight Recommendation Systems in Production How recommendation systems become production decisioning systems through signals, ranking, constraints, feedback loops, and experimentation Readers have engaged with this next