Concepedia

Multi-step Returns

Multi-Step Returns is a value function estimation method in reinforcement learning that generalizes the one-step return by incorporating discounted future rewards over a fixed number of steps, thereby balancing bias and variance in temporal difference learning.

4

Publications

409

Citations

11

Authors

4

Institutions

Publications per year

2017–2026

2

Variants

N-step Returns

Parents

Publications

4

1–4 of 4

Authors

11

  1. MM
    Melanie Marks

    Longwood University, United States

    2 publications · 366 citations

    2

    h-index

  2. RC
    Rachel Croson

    University of Pennsylvania, United States

    2 publications · 366 citations

    2

    h-index

  3. DP
    Doina Precup

    McGill University, Canada

    1 publications · 14 citations

    1

    h-index

  4. PV
    Peter Vrancx

    Vrije Universiteit Brussel, Belgium

    1 publications · 14 citations

    1

    h-index

  5. AN
    Ann Nowé

    Vrije Universiteit Brussel, Belgium

    1 publications · 14 citations

    1

    h-index

1–5 of 11

Institutions

4

  1. Longwood University

    Farmville, United States

    2 h-index

    2 publications

    366 citations

  2. University of Pennsylvania

    Philadelphia, United States

    2 h-index

    2 publications

    366 citations

  3. Vrije Universiteit Brussel

    Brussels, Belgium

    1 h-index

    2 publications

    28 citations

  4. McGill University

    Montreal, Canada

    1 h-index

    2 publications

    28 citations

1–4 of 4

Venues

  1. Experimental Economics

    2 publications · 366 citations

    2

    h-index

1–1