Time-critical and confidence-based abstraction dropping methods

Robin Schmöcker, Lennart Kampmann, Alexander Dockhorn
7/3/2025
cs.AI

No official code found. Best implementation: codebro634/AbsDropping (14 stars, MIT, last commit 2025-11-18).

Abstract

One paradigm of Monte Carlo Tree Search (MCTS) improvements is to build and use state and/or action abstractions during the tree search. Non-exact abstractions, however, introduce an approximation error making convergence to the optimal action in the abstract space impossible. Hence, as proposed as a component of Elastic Monte Carlo Tree Search by Xu et al., abstraction algorithms should eventually drop the abstraction. In this paper, we propose two novel abstraction dropping schemes, namely OGA-IAAD and OGA-CAD which can yield clear performance improvements whilst being safe in the sense that the dropping never causes any notable performance degradations contrary to Xu's dropping method. OGA-IAAD is designed for time critical settings while OGA-CAD is designed to improve the MCTS performance with the same number of iterations.

View on arXivView PDF

Code Implementations(1)

codebro634/AbsDroppingCommunity72%
MIT

Code and supplementary materials for the paper "Time-critical and confidence-based abstraction dropping methods"

140C++, CMakeFeb 13, 202511 months agoMIT

Similar Papers

Accelerating Monte Carlo Tree Search with Probability Tree State Abstraction

77% match1 repo172023

Towards Tackling MaxSAT by Combining Nested Monte Carlo with Local Search

74% match2023

Monte Carlo Tree Search for Asymmetric Trees

73% match1 repo132018

Probe-and-Resume: When and How to Evaluate Agent Execution

73% match2026

Explicit Dropout: Deterministic Regularization for Transformer Architectures

72% match2026

3D-Prover: Diversity Driven Theorem Proving With Determinantal Point Processes

72% match1 repo12024

Cite this paper

@article{schmcker2025timecritical,
  title  = {Time-critical and confidence-based abstraction dropping methods},
  author = {Robin Schmöcker and Lennart Kampmann and Alexander Dockhorn},
  year   = {2025},
  eprint = {2507.02703},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2507.02703v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2507.02703.svg)](https://codeofpaper.com/paper/2507.02703)

Discussion