AI Navigate

マスクなしポリシー勾配アルゴリズムにおける有効行動抑制の克服

arXiv cs.LG / 2026/3/11

Ideas & Deep AnalysisModels & Research

要点

  • 本論文は、マスクなしポリシー勾配強化学習において、未訪問状態での有効な行動が、訪問済み状態での無効行動による勾配の影響で体系的に抑制されるという失敗モードを特定している。
  • 著者らは、共有特徴を持つソフトマックスポリシーに対し、訪問済み状態では無効だが未訪問状態で有効な行動の確率が、パラメータ共有とソフトマックスの特性により指数関数的に抑制されることを数学的に証明している。
  • エントロピー正則化は、有効行動の保護とサンプル効率のバランスを取ることでこの抑制を部分的に緩和するが、行動マスキングがこの問題を完全に解消することを示している。
  • CraftaxやMiniHackなどでの実験的検証により、理論的な指数関数的抑制が確認され、フィージビリティ分類が実際にはオラクルマスクの代替として機能することを実証している。
  • 本研究は、行動マスキングがポリシー勾配学習を改善し抑制を防ぐ仕組みを理解する上で進展をもたらし、状態依存の有効行動処理に対してより効果的なアプローチを強化学習アルゴリズムに提供する。

Computer Science > Machine Learning

arXiv:2603.09090 (cs)
[Submitted on 10 Mar 2026]

Title:Overcoming Valid Action Suppression in Unmasked Policy Gradient Algorithms

View a PDF of the paper titled Overcoming Valid Action Suppression in Unmasked Policy Gradient Algorithms, by Renos Zabounidis and 5 other authors
View PDF HTML (experimental)
Abstract:In reinforcement learning environments with state-dependent action validity, action masking consistently outperforms penalty-based handling of invalid actions, yet existing theory only shows that masking preserves the policy gradient theorem. We identify a distinct failure mode of unmasked training: it systematically suppresses valid actions at states the agent has not yet visited. This occurs because gradients pushing down invalid actions at visited states propagate through shared network parameters to unvisited states where those actions are valid. We prove that for softmax policies with shared features, when an action is invalid at visited states but valid at an unvisited state $s^*$, the probability $\pi(a \mid s^*)$ is bounded by exponential decay due to parameter sharing and the zero-sum identity of softmax logits. This bound reveals that entropy regularization trades off between protecting valid actions and sample efficiency, a tradeoff that masking eliminates. We validate empirically that deep networks exhibit the feature alignment condition required for suppression, and experiments on Craftax, Craftax-Classic, and MiniHack confirm the predicted exponential suppression and demonstrate that feasibility classification enables deployment without oracle masks.
Subjects: Machine Learning (cs.LG)
Cite as: arXiv:2603.09090 [cs.LG]
  (or arXiv:2603.09090v1 [cs.LG] for this version)
  https://doi.org/10.48550/arXiv.2603.09090
Focus to learn more
arXiv-issued DOI via DataCite

Submission history

From: Renos Zabounidis [view email]
[v1] Tue, 10 Mar 2026 02:07:37 UTC (8,026 KB)
Full-text links:

Access Paper:

Current browse context:
cs.LG
< prev   |   next >
Change to browse by:
cs

References & Citations

export BibTeX citation Loading...

BibTeX formatted citation

×
Data provided by:

Bookmark

BibSonomy logo Reddit logo
Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
Links to Code Toggle
Papers with Code (What is Papers with Code?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos

Demos

Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers

Recommenders and Search Tools

Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
IArxiv recommender toggle
IArxiv Recommender (What is IArxiv?)
About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.