Methods › Reinforcement Learning › Value Function Estimation › Retrace

Retrace

31 papers tagged archive 2025-07-28

Introduced by Rémi Munos et al. in Safe and Efficient Off-Policy Reinforcement Learning

archive 2025-07-28 Description, source and code snippet are the archive's method entry.

Retrace is an off-policy Q-value estimation algorithm which has guaranteed convergence for a target and behaviour policy (π, β). With off-policy rollout for TD learning, we must use importance sampling for the update:

ΔQⁱᵐᵖ(Sₜ, Aₜ) = γᵗ∏_(1≤τ≤t)(π(A_τ|S_τ))/(β(A_τ|S_τ))δₜ

This product term can lead to high variance, so Retrace modifies ΔQ to have importance weights truncated by no more than a constant c:

ΔQⁱᵐᵖ(Sₜ, Aₜ) = γᵗ∏_(1≤τ≤t)min(c, (π(A_τ|S_τ))/(β(A_τ|S_τ)))δₜ

PaperSource

Papers archive 2025-07-28

30 shown of 31, newest first. Repository counts are the archive's code-links table. A Syntology line states what Syntology ran from that paper's harvested code; it is per sample and not a correctness claim.

Tasks archive 2025-07-28

20 shown of 38 tasks the archive attaches to papers tagged with this method, by distinct papers. A task without a page in the catalog is plain text.

TaskPapers
Reinforcement Learning (RL)13
Reinforcement Learning12
reinforcement-learning12
Deep Reinforcement Learning9
Face Anti-Spoofing3
Face Recognition3
Atari Games2
General Classification2
Partially Observable Reinforcement Learning2
Problem Decomposition2
Time Series2
Time Series Analysis2
Automatic Speech Recognition1
Automatic Speech Recognition (ASR)1
Benchmarking1
Binary Classification1
Continuous Control1
Data Augmentation1
Decision Making1
Deep Learning1

Usage over time archive 2025-07-28

Papers per year tagged with Retrace: 2016 to 2025, peak 7 7 0 2016: 2 papers 2016 2017: 1 paper 2017 2018: 2 papers 2018 2019: 4 papers 2019 2020: 1 paper 2020 2021: 5 papers 2021 2022: 7 papers 2022 2023: 3 papers 2023 2024: 4 papers 2024 2025: 2 papers 2025
Papers per year the archive tags with this method, by the paper's archive date (31 dated). Bars are counts, not a trend claim.

Components: the archive holds no method-to-method composition, so PwC's Components table cannot be rebuilt; the Papers list carries no Results column for the same reason (the archive does not join its leaderboard rows to method tags).

Categories archive 2025-07-28

Value Function Estimation

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections