Back to Square One: Superhuman Performance in Chutes and Ladders Through Deep Neural Networks and Tree Search
arXiv:2104.00698
Abstract
We present AlphaChute: a state-of-the-art algorithm that achieves superhuman performance in the ancient game of Chutes and Ladders. We prove that our algorithm converges to the Nash equilibrium in constant time, and therefore is -- to the best of our knowledge -- the first such formal solution to this game. Surprisingly, despite all this, our implementation of AlphaChute remains relatively straightforward due to domain-specific adaptations. We provide the source code for AlphaChute here in our Appendix.
References in corpus (18)
- Distilling the Knowledge in a Neural Network
- Improving neural networks by preventing co-adaptation of feature detectors
- A Simple Way to Initialize Recurrent Networks of Rectified Linear Units
- Similarity of Neural Network Representations Revisited
- Regularizing Neural Networks by Penalizing Confident Output Distributions
- Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer
- Distilling a Neural Network Into a Soft Decision Tree
- Using Fast Weights to Attend to the Recent Past
- Conditional Restricted Boltzmann Machines for Structured Output Prediction
- Deep Lambertian Networks
- Deep Mixtures of Factor Analysers
- Analyzing and Improving Representations with the Soft Nearest Neighbor Loss
- Discovering Multiple Constraints that are Frequently Approximately Satisfied
- Unsupervised part representation by Flow Capsules
- Products of Hidden Markov Models: It Takes N>1 to Tango
- Cerberus: A Multi-headed Derenderer
- Efficient Parametric Projection Pursuit Density Estimation
- Teaching with Commentaries