# Solving the Rubik's Cube Without Human Knowledge

arXiv: https://arxiv.org/abs/1805.07470 · HTML: https://mlsanity.com/paper/1805.07470/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 18 May 2018.

## Built on by (most cited first, 15 of 15)

- [Jump-Start Reinforcement Learning](https://mlsanity.com/paper/2204.02372/) (2022)
- [Puzzle Solving using Reasoning of Large Language Models: A Survey](https://mlsanity.com/paper/2402.11291/) (2024)
- [AlphaD3M: Machine Learning Pipeline Synthesis](https://mlsanity.com/paper/2111.02508/) (2021)
- [Imperative Learning: A Self-supervised Neuro-Symbolic Learning Framework for Robot Autonomy](https://mlsanity.com/paper/2406.16087/) (2024)
- [Puzzle Solving without Search or Human Knowledge: An Unnatural Language Approach](https://mlsanity.com/paper/2109.02797/) (2021)
- [Graph Neural Networks for Image Classification and Reinforcement Learning using Graph representations](https://mlsanity.com/paper/2203.03457/) (2022)
- [Learning Shortest Paths with Generative Flow Networks](https://mlsanity.com/paper/2603.01786/) (2026)
- [On Solving the Rubik's Cube with Domain-Independent Planners Using Standard Representations](https://mlsanity.com/paper/2307.13552/) (2023)
- [Solving a Rubik's Cube Using its Local Graph Structure](https://mlsanity.com/paper/2408.07945/) (2024)
- [Node Classification and Search on the Rubik's Cube Graph with GNNs](https://mlsanity.com/paper/2501.18580/) (2025)
- [Diffusion Models for Cayley Graphs](https://mlsanity.com/paper/2503.05558/) (2025)
- [Towards Learning Rubik's Cube with N-tuple-based Reinforcement Learning](https://mlsanity.com/paper/2301.12167/) (2023)
- [Improved Tree Search for Automatic Program Synthesis](https://mlsanity.com/paper/2303.07166/) (2023)
- [Learning Empirically Admissible Neural Heuristics for Combinatorial Search](https://mlsanity.com/paper/2606.04860/) (2026)
- [Playing to Par: Reinforcement Learning for Provably Optimal Quadrilateral Block Decompositions](https://mlsanity.com/paper/2609.32146/) (2026)

## Abstract (arXiv)

A generally intelligent agent must be able to teach itself how to solve problems in complex domains with minimal human supervision. Recently, deep reinforcement learning algorithms combined with self-play have achieved superhuman proficiency in Go, Chess, and Shogi without human data or domain knowledge. In these environments, a reward is always received at the end of the game, however, for many combinatorial optimization environments, rewards are sparse and episodes are not guaranteed to terminate. We introduce Autodidactic Iteration: a novel reinforcement learning algorithm that is able to teach itself how to solve the Rubik's Cube with no human assistance. Our algorithm is able to solve 100% of randomly scrambled cubes while achieving a median solve length of 30 moves -- less than or equal to solvers that employ human domain knowledge.