• Publications
  • Influence
Learning to Play the Chess Variant Crazyhouse Above World Champion Level With Deep Neural Networks and Human Data
TLDR
Improvements include modifications in the neural network design and training configuration, the introduction of a data normalization step and a more sample efficient Monte-Carlo tree search which has a lower chance to blunder. Expand
Monte-Carlo Graph Search for AlphaZero
TLDR
A new, improved search algorithm for AlphaZero is introduced which generalizes the search tree to a directed acyclic graph, which enables information flow across different subtrees and greatly reduces memory consumption. Expand
Distributed Methods for Reinforcement Learning Survey
  • Johannes Czech
  • Computer Science
  • Reinforcement Learning Algorithms: Analysis and…
  • 2021
TLDR
This survey presents several distributed methods including multi-agent schemes, synchronous and asynchronous parallel systems, as well as population-based approaches to address the issue of high computational requirements for reinforcement learning. Expand
Generative Adversarial Neural Cellular Automata
TLDR
GANCA, a novel algorithm that combines Neural Cellular Automata with Generative Adversarial Networks, allows for more generalization through adversarial training, and shows that a single model is capable of learning several images when presented with different inputs. Expand
Improving AlphaZero Using Monte-Carlo Graph Search
TLDR
This work improves the search algorithm for AlphaZero by generalizing the search tree to a directed acyclic graph, which enables information flow across different subtrees and greatly reduces memory consumption. Expand