Concepedia

Off-policy Evaluation

Off-Policy Evaluation is a methodology in recommender systems research that enables the estimation of a new policy's performance using historical data collected under a different logging policy, thereby allowing unbiased evaluation without requiring new real-time interactions.

121

Publications

7.7K

Citations

308

Authors

97

Institutions

Publications per year

2017–2026

93

Variants

Ope

Parents

Publications

121

Soft Actor-Critic Algorithms and Applications

Tuomas Haarnoja, Aurick Zhou, Kristian Hartikainen et al. · arXiv (Cornell University) · 2018 · 2K citations · Full text

, , +12

Q-Learning Algorithms: A Comprehensive Classification and Applications

Beakcheol Jang, Myeonghwi Kim, Gaspard Harerimana et al. · IEEE Access · 2019 · 604 citations · Full text

, , +14

Eligibility Traces for Off-Policy Policy Evaluation

Doina Precup, Richard S. Sutton, Satinder Singh · Scholarworks (University of Massachusetts Amherst) · 2000 · 462 citations · Full text

, , +13

Horde: a scalable real-time architecture for learning knowledge from unsupervised sensorimotor interaction

Richard S. Sutton, Joseph Modayil, Michael Delp et al. · 2011 · 305 citations

, , +14

Off-Policy Temporal Difference Learning with Function Approximation

Doina Precup, Richard S. Sutton, Sanjoy Dasgupta · 2001 · 254 citations

, , +13

1–5 of 121

Authors

308

  1. ON
    Ofir Nachum

    Google (United States), United States

    8 publications · 239 citations

    8

    h-index

  2. MU
    Masatoshi Uehara

    Cornell University, United States

    7 publications · 224 citations

    7

    h-index

  3. MG
    Mohammad Ghavamzadeh

    Adobe Systems (United States), United States

    6 publications · 479 citations

    6

    h-index

  4. NK
    Nathan Kallus

    Cornell University, United States

    6 publications · 206 citations

    6

    h-index

  5. PS
    Philip S. Thomas

    University of Massachusetts Amherst, United States

    6 publications · 408 citations

    6

    h-index

1–5 of 308

Institutions

97

  1. Google (United States)

    Mountain View, United States

    13 h-index

    61 publications

    2K citations

  2. Cornell University

    Ithaca, United States

    11 h-index

    16 publications

    442 citations

  3. Stanford University

    Stanford, United States

    8 h-index

    14 publications

    317 citations

  4. Carnegie Mellon University

    Pittsburgh, United States

    7 h-index

    16 publications

    434 citations

  5. University of Alberta

    Edmonton, Canada

    7 h-index

    16 publications

    2K citations

  6. University of California, Berkeley

    Berkeley, United States

    7 h-index

    11 publications

    221 citations

  7. University of Illinois Urbana-Champaign

    Urbana, United States

    5 h-index

    7 publications

    177 citations

  8. Microsoft (United States)

    Redmond, United States

    5 h-index

    12 publications

    412 citations

1–8 of 97

Venues

  1. arXiv (Cornell University)

    Repository · 63 publications · 4K citations

    26

    h-index

  2. Proceedings of the AAAI Conference on Artificial Intelligence

    Conference · 7 publications · 263 citations

    7

    h-index

  3. International Conference on Machine Learning

    Conference · 4 publications · 62 citations

    4

    h-index

  4. Neurocomputing

    3 publications · 68 citations

    3

    h-index

  5. Operations Research

    2 publications · 80 citations

    2

    h-index

1–5