Methods › General › Regularization › Target Policy Smoothing

Target Policy Smoothing

116 papers tagged archive 2025-07-28

Introduced by Scott Fujimoto et al. in Addressing Function Approximation Error in Actor-Critic Methods

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

Target Policy Smoothing is a regularization strategy for the value function in reinforcement learning. Deterministic policies can overfit to narrow peaks in the value estimate, making them highly susceptible to functional approximation error, increasing the variance of the target. To reduce this variance, target policy smoothing adds a small amount of random noise to the target policy and averages over mini-batches - approximating a SARSA-like expectation/integral.

The modified target update is:

y = r + γQ_(θ′)(s′, π_(θ′)(s′) + ϵ)

ϵ∼clip(𝒩(0, σ), -c, c )

where the added noise is clipped to keep the target close to the original action. The outcome is an algorithm reminiscent of Expected SARSA, where the value estimate is instead learned off-policy and the noise added to the target policy is chosen independently of the exploration policy. The value estimate learned is with respect to a noisy policy defined by the parameter σ.

PaperSourceSee Code · sfujim/TD3

In Syntology for a different name, 'policy smoothing', about another paper, arXiv:2106.11420: verified reference implementation at level V3; retrieve with get_reference_implementation(name='policy smoothing'). The lookup matched a different name in Syntology's graph, 'policy smoothing' (resolution: token_match); this answer is about that name. About another paper: this answer is about arXiv:2106.11420, while the archive credits arXiv:1802.09477.

Origin paper of this answer, arXiv:2106.11420 (its page on Syntology).

Syntology's have check for this name (free, it answers coverage only), read 2026-10-05; ask it again live. A level is the served verification level of that one answer; nothing here is summed across methods.

What each level means, quoted from Syntology's server card: V0 “imports”; V1 “executes against a synthesized fixture and returns a value of the recorded type”; V2 “two implementations GENERATED from the paper's signature and spec agree numerically (1e-6) on fixture inputs” “neither of the two ever read the authors' own code, so V2 measures how unambiguously the paper specifies the method, not fidelity to the authors' implementation (measured 2026-09-11, commit 75eaa77: cross_checked_by is a second LLM on 2,628 of 2,831 served samples)”; V3 “paper-derived property tests pass on both of those same two generations.” The card adds: “No rung on this ladder compares the served code against the authors' implementation, so treat every level as evidence about runnability and internal agreement, not about matching the paper's results.” A level quoted in two parts is one definition the card writes as two clauses joined by a double hyphen, each clause quoted on its own (how these levels are quoted).

Papers archive 2025-07-28

archive papers tagged: 116 · with a code link: 45 · where Syntology ran a sample: 18 (17 with a run with no instrument failure, 1 where every run was a failure of Syntology's instrument) Syntology

Show: all tagged papersonly where code ran (18 of 116 tagged: 17 with a run with no instrument failure, 1 where every run was a failure of Syntology's instrument)

Page 1 of 2: papers 1 to 100 of 116, newest first by the archive's date (ties by slug), in archive order.

Repository counts are the archive's code-links table. A Syntology line states what Syntology ran from that paper's harvested code, as “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified”; the instrument figure counts failures of Syntology's instrument, not of the code. It is per sample and not a correctness claim. When the archive marks a repository official for the paper, the line starts with that repository's state (the archive's flag, not a verdict on who wrote the code; “community repositories only” when every sample that ran came from a community repository, “official: no sample here; runs from other or unrecorded repositories” when some came from a repository the paper names or has in its text, or from none recorded); hover it for the repositories the samples that ran came from.

Tasks archive 2025-07-28

20 shown of 69 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)67
reinforcement-learning54
Deep Reinforcement Learning53
Reinforcement Learning46
Continuous Control30
continuous-control29
MuJoCo25
Q-Learning11
OpenAI Gym9
Decision Making8
Autonomous Driving6
Offline RL5
Benchmarking4
Meta-Learning4
Policy Gradient Methods4
D4RL3
GPU3
Scheduling3
Atari Games2
Computational Efficiency2

Usage over time archive 2025-07-28

Papers per year tagged with Target Policy Smoothing: 2018 to 2025, peak 25 25 0 2018: 1 paper 2018 2019: 6 papers 2019 2020: 18 papers 2020 2021: 21 papers 2021 2022: 10 papers 2022 2023: 25 papers 2023 2024: 25 papers 2024 2025: 10 papers 2025
Papers per year the archive tags with this method, by the paper's archive date (116 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

Regularization

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