• Home
  • Search
  • Dynamic Reinforcement Learning for Optimal Go AI Training: Adaptive Adjustment and Optimization
  • Cite Icon7
  • https://doi.org/10.1109/tce.2024.3487141Copy DOI Icon

Dynamic Reinforcement Learning for Optimal Go AI Training: Adaptive Adjustment and Optimization

Show More
  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Go is a popular strategy game today, but due to its large search space and task complexity, ensuring stable AI implementation is challenging. Specifically, Go AI training requires setting a fixed optimal learning rate and schedule, which demands significant TPU and GPU resources. To facilitate Go-AI learning, this research explores adaptive adjustment and optimization techniques for dynamic reinforcement learning neural networks. First, we introduce a dynamic batch size technique that adjusts data volume at each training phase and incorporates dynamic network structure search, considering the number of network layers and residual blocks. Second, we propose a dynamic network topology that automatically modifies the learning rate based on the training batch size at various training phases. Our approach outperforms the baseline in terms of stability and model convergence speed. In 100 games, the Go-AI model achieved a 100% victory rate below the 7th rank and a 98% win rate at the 9th rank and higher.

Similar Papers
  • Research Article
  • Citations179

Agile convolutional neural network for pulmonary nodule classification using CT images.

  • Feb 23, 2018
  • International Journal of Computer Assisted Radiology and Surgery
  • Xinzhuo Zhao +5
  • Research Article
  • Citations19

A new perspective for understanding generalization gap of deep neural networks trained with large batch sizes

  • Nov 24, 2022
  • Applied Intelligence
  • Oyebade K Oyedotun +2
  • PDF
  • Research Article
  • Citations1

DyLFG: A Dynamic Network Learning Framework Based on Geometry

  • Nov 30, 2023
  • Entropy
  • Wei Wu +1
  • PDF
  • Conference Article
  • Citations21

Online evolutionary batch size orchestration for scheduling deep learning workloads in GPU clusters

  • Nov 13, 2021
  • Zhengda Bian +3
  • Research Article
  • Citations4

The Limiting Dynamics of SGD: Modified Loss, Phase-Space Oscillations, and Anomalous Diffusion.

  • Dec 12, 2023
  • Neural Computation
  • Daniel Kunin +6
  • Supplementary Content

TerraSAR-X Ice/Non-Ice Mapping of the Mackenzie River Using a Convolutional Neural Network

  • Jan 01, 2019
  • elib (German Aerospace Center)
  • Felix Reuß +2
  • Research Article
  • Citations2

Dynamic network link prediction based on learning continuous time events

  • Feb 01, 2023
  • SCIENTIA SINICA Informationis
  • 忠明 韩 +4
  • Research Article

A Compositional Framework for Neural Network Modeling Using Typed Petri Nets

  • Feb 28, 2026
  • Transactions of the Japanese Society for Artificial Intelligence
  • Yiyang Jia +3
  • Research Article
  • Citations93

Adaptive Batch Size for Federated Learning in Resource-Constrained Edge Computing

  • Jan 01, 2023
  • IEEE Transactions on Mobile Computing
  • Zhenguo Ma +5
  • Research Article
  • Citations11

Differentially private graph neural networks for graph classification and its adaptive optimization

  • Nov 19, 2024
  • Expert Systems With Applications
  • Yong Li +4
  • Conference Article
  • Citations6

Simultaneous Localization and Anomaly Detection from First-Person Video Data through a Coupled Dynamic Bayesian Network Model

  • Nov 29, 2022
  • Giulia Slavic +4
  • Conference Article
  • Citations1

Dynamic network representation learning based on temporal neighborhood aggregation

  • Oct 13, 2022
  • Fengzhe Zhong +2
  • Conference Article
  • Citations6

Evolutionary learning of policies for MCTS simulations

  • May 29, 2012
  • James Pettit +1
  • Research Article
  • Citations22

Multi-objective ensemble deep learning using electronic health records to predict outcomes after lung cancer radiotherapy

  • Dec 13, 2019
  • Physics in Medicine & Biology
  • Rongfang Wang +5
  • Conference Article
  • Citations2

Robustifying DARTS by Eliminating Information Bypass Leakage via Explicit Sparse Regularization

  • Dec 01, 2021
  • Jiuling Zhang +1
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.