# Biasing MCTS with Features for General Games

arXiv: https://arxiv.org/abs/1903.08942 · HTML: https://mlsanity.com/paper/1903.08942/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 21 Mar 2019.

## Built on by (most cited first, 4 of 4)

- [Collaborative Agent Gameplay in the Pandemic Board Game](https://mlsanity.com/paper/2103.11388/) (2021)
- [Spatial State-Action Features for General Games](https://mlsanity.com/paper/2201.06401/) (2022)
- [Contrastive Sparse Autoencoders for Interpreting Planning of Chess-Playing Agents](https://mlsanity.com/paper/2406.04028/) (2024)
- [Optimised Playout Implementations for the Ludii General Game System](https://mlsanity.com/paper/2111.02839/) (2021)

## Abstract (arXiv)

This paper proposes using a linear function approximator, rather than a deep neural network (DNN), to bias a Monte Carlo tree search (MCTS) player for general games. This is unlikely to match the potential raw playing strength of DNNs, but has advantages in terms of generality, interpretability and resources (time and hardware) required for training. Features describing local patterns are used as inputs. The features are formulated in such a way that they are easily interpretable and applicable to a wide range of general games, and might encode simple local strategies. We gradually create new features during the same self-play training process used to learn feature weights. We evaluate the playing strength of an MCTS player biased by learnt features against a standard upper confidence bounds for trees (UCT) player in multiple different board games, and demonstrate significantly improved playing strength in the majority of them after a small number of self-play training games.