Skip to content

QSeaBattle

QSeaBattle is a simulation and machine-learning project for exploring classical, learned, and non-classical strategies in distributed guessing games.

The project grew out of a series of investigations into random access codes, nonlocal correlations, trainable assisted strategies, and the geometry behind quantum advantage.

  • Read The Story

    The ideas behind QSeaBattle are explained in a series of articles on
    The Armchair Quantum Physicist.

    Read the articles →

  • Explore The Code

    Browse the source code, notebooks, tests, and simulations used in the project.

    Open the GitHub repository →

  • Technical Documentation

    Explore the algorithms, players, models, utilities, and generated API documentation.

    Browse the documentation →

Tutorials

The project provides tutorials that show how the game, tournaments and models can be used.

  1. A quick start guide showing how to run the provided players in a tournament

  2. Demonstration of a neural-net player trained to imitate the majority algorithm

  3. Self-training of a neural-net-based player

  4. Imitation training of a player using post-quantum resources with the "linear" model

  5. Imitation training of a player using post-quantum resources with the "pyramid" model

The Article Series

The project is accompanied by five articles that develop the ideas step by step.

  1. The Best Bit Alice Can Send
    Why the most informative message isn’t the winning one.

  2. Beating the Odds in a Guessing Game — with a Single Quantum Box
    Alice sends Bob one bit and tells him nothing new — yet he guesses right more often than he should. A small quantum advantage, in the simplest game there is.

  3. Quantum Mechanics Should Make This More Complicated. It Doesn’t.
    A guessing game becomes geometrically more and more complicated — until we add quantum mechanics.

  4. Why Isn’t Nature More Quantum? (to be published)
    Quantum physics draws a boundary. Information theory approaches the same limit from a completely different direction.

  5. Is quantum mechanics the optimal balance between what we can know and what we can do? (to be published)
    Quantum theory achieves in some sense an optimal balance of allowed states and dynamics.

What You Will Find Here

The documentation covers:

  • the game and tournament framework
  • classical baseline strategies
  • majority and assisted players
  • PR-assisted strategies
  • trainable assisted models
  • linear and pyramidal architectures
  • dataset generation and conversion
  • imitation-learning utilities
  • analysis and reference-performance tools

If you are mainly interested in the implementation, start with the Algorithms and Players sections.

If you want to understand the motivation and the physics behind the project, start with the article series above.

Repository Structure

  • docs/ — Documentation source
  • notebooks/ — Tutorials and experiments
  • presentation/ — Presentation material
  • src/ — QSeaBattle source code
  • tests/ — Automated tests
  • tools/ — Documentation and maintenance tools