Method

Multi-armed bandit

Adaptive allocation with Thompson sampling, epsilon-greedy and contextual bandits.

40 cases from 25 companies apply Multi-armed bandit. 18 were checked in an online A/B test; 5 report results in production. Includes variants: Thompson sampling.

Typical tasks: Recommend to new users and new items, Optimize decisions, Avoid false wins in A/B tests, Rank search results and 12 more. Companies: DoorDash, Target, Netflix, Instacart, Lyft, Ramp and 19 more. Often combined with: Thompson sampling, A/B test, Gradient boosting, Rule-based heuristics, Doubly robust estimation and 66 more.

Crowdsourced taxonomy verification: A feedback-driven framework for refining knowledge graph relationships via online search interactionsfood-delivery2026-07
Online Learning for Cost-Efficient LLM Routingfintech2026-07
What If Your Page Layout Could Learn? Optimizing with Banditse-commerce2026-05
Why we use separate tech stacks for personalization and experimentationmedia-streaming2026-01
Accelerating experimentation at DoorDash with a multi-armed bandit platformfood-delivery2025-12
Chasing All the Hares: Managing Different Content Types in One Recommender Systembanking2025-05
How machine learning supercharged our revenue by millions of dollarsedtech2025-03
Next Best Action platform: democratizing personalization with contextual banditsmarketplace2024-12
Transforming Location Retrieval at Airbnb: A Journey from Heuristics to Reinforcement Learningtravel2024-11
Identifying Top-Scoring Arms in Ranking Bandits With Linear Payoffs in Real-Timetravel2024-11
How The New York Times Incorporates Editorial Judgement in Algorithms to Curate Home Screen Contentmedia-streaming2024-10
Recommending for Long-Term Member Satisfaction at Netflixmedia-streaming2024-08
Contextual Offer Recommendations Engine at Targetretail2024-08
Bandits for Marketing Optimizationfood-delivery2024-06
Make Better Decisions by Embracing Uncertaintyfintech2024-06
Warming up a Cold Usermedia-streaming2024-05
Technology Trends in RecSysgeneral2024-05
Matching LinkedIn members with the right Premium productssocial2024-05
Lyft’s Reinforcement Learning Platformmobility2024-03
How The New York Times Cooking Team Makes Personalized Recipe Recommendationsmedia-streaming2023-10
End-to-end Keyword Bidding for Apple Search Adsadtech2023-09
Using Contextual Bandit models in large action spaces at Instacarte-commerce2023-06
Real-Time Personalization Using Microservicesretail2023-06
Multi-Armed Bandits in Yandex Lavkafood-delivery2023-05
RLHF for retrieval based NLP systemgeneral2023-05
A Review of Multi-Armed Bandits Applications at Lyftmobility2023-01
Selecting the Best Image for Each Merchant Using Exploration and Machine Learningfood-delivery2023-01
Lessons from A/B Testing on Bandit Subjectsadtech2022-12
How to Optimise Rankings with Cascade Banditstravel2022-12
Personalizing Recommendations for a Learning Usere-commerce2022-12
Explore-exploit dilemma in Ranking modeltravel2022-11
Homepage Recommendation with Exploitation and Explorationfood-delivery2022-10
Multi-Objective Ranking for Promoted Auction Itemsmarketplace2022-09
Reinforcement Learning for Budget Constrained Recommendationsmedia-streaming2022-08
How Bandit Testing Improves A/B Tests at Wealthfrontfintech2022-06
Challenges in Experimentationmobility2022-04
Going from 0 to 1 modeling User Preferences for Personalized Recommendationsmarketplace2022-03
Using a Multi-Armed Bandit with Thompson Sampling to Identify Responsive Dashersfood-delivery2022-03
Using Bayesian optimization for balancing metrics in recommendation systemssocial2022-02
Experimentation is a major focus of Data Science across Netflixmedia-streaming2022-01