Zzong's Notes

Home

❯

papers

❯

_archive

❯

Making contextual decisions with low technical debt

Making contextual decisions with low technical debt

2026년 8월 29일1 min read

paper link: https://arxiv.org/abs/1606.03966

Contextual Bandit pipeline 을 구성할 때 어떻게 하면 기술 부채가 적어지는지에 대한 내용을 다루는 것 같음


함께 보면 좋은 글

Mortal Multi-Armed Bandits

Mortal MAB, Multi-Armed Bandit Related Mortal Multi Armed Bandit (2008) References papers.nips.cc/paper/2008/file/788d986905533aba051261497ecffcbb-Paper.pdf.

An Asymptotically Optimal Primal-Dual Incremental Algorithm for Contextual Linear Bandits

Links Paper link Abstract optimism principle 에 기반한 알고리즘은 문제에 대한 구조를 exploit 하는데 실패하여 점근적으로 suboptimal 결과를 보임 context 분포와 exploration policy 가 나눠지도록 (decoupled) regret lower...

Top-K Contextual Bandits with Equity of Exposure

Abstract Probability Ranking Principle 에 의하면 top-K items 을 greedy 하게 rank 하는것이 optimal 함 대신 Introduction This work investigates how the “equity of exposure” principle can be...

Contextual Combinatorial Bandit and its Application on Diversified Online Recommendation

References www.chenshouyuan.com/papers/sdm14.pdf .

Exploring compact reinforcement-learning representations with linear regression

paper Link: arxiv.org/pdf/1205.2606.pdf KWIK Linear Regression KWIK (Knows What It Knows) is a framework for studying supervised learning algorithms and was designed to unify...

Recommendations with Negative Feedback via Pairwise Deep Reinforcement Learning

Discussion Q-learning based offline 학습 방식이고, 모델도 무겁고.

Deep Exploration via Bootstrapped DQN

Abstract 해결하려는 문제: 강화학습에서의 효율적인 exploration Randomized value functions offer a promising approach to efficient exploration with generalization, but existing algorithms are not...

A Survey on Reinforcement Learning for Recommender Systems

References paper link: arxiv.org/pdf/2109.10665.pdf Reinforcement Learning based Recommender Systems: A Survey .

Reinforcement Learning for Slate-based Recommender Systems - A Tractable Decomposition and Practical Methodology

Empirical Evaluation: Live Experiments YouTube 에 SARSA-TS 알고리즘을 실험 candidate -> ranker 를 거치게 되는데, ranker 의 scoring 함수에서 사용하는 myopic(근시안적) engagement 측정값을 LTV estimate 로 변경함...

Deep reinforcement learning for search, recommendation, and online advertising - a survey

Paper link arxiv.org/abs/1812.07127 .