sudoku

This position paper argues that neural solvers for constraint satisfaction problems must prioritize symbolic integration to ensure provable correctness, especially under distribution shifts, using Sudoku as a key example.

Introduces Simplax, an exact Dirichlet-categorical augmentation for uniform discrete diffusion that improves reverse sampling and generative quality on text and Sudoku tasks.

A researcher's blog post describes turning constraint models into playable puzzle games, based on his paper on scaling Sudoku as a constraint problem. It shares MiniZinc models, a repository of 434,201 Sudoku instances, and playable versions of nine puzzle games.

This paper proposes an oscillatory neural network (ONN) based solver for Sudoku puzzles by formulating them as graph coloring problems, achieving high accuracy on 4x4 and 9x9 puzzles.

This paper proposes a neuro-symbolic approach that integrates a MaxSAT oracle as a consistency validator to guide Vision-Language Models (VLMs) in solving Sudoku puzzles, improving logical consistency and the number of solved instances.

Introduces the Energy-guided Recursive Model (ERM), which uses Hopfield energies to guide selection among recursive reasoning trajectories, achieving state-of-the-art performance on Sudoku, Pencil Puzzle Bench, and Maze tasks.

This paper presents a 2048-spin bulk acoustic wave Ising machine using microwave delay lines, achieving all-to-all connectivity and solving MAX-CUT, number partitioning, and Sudoku problems with higher thermal stability and performance compared to state-of-the-art coherent Ising machines.

A developer created Puzzle Lair, a free logic puzzle site featuring Sudoku, Kakuro, Nonograms and more, with no ads and a one-time unlock for full catalogs.

Proposes DiBS, a diffusion model-guided approach for branch selection in exact Sudoku solvers that reduces search cost without sacrificing completeness, supported by theoretical proof and empirical results on the Royle 17-clue benchmark.

This paper introduces a method to make Tiny Recursive Models stochastic at test time by adding Gaussian noise and running parallel rollouts, achieving dramatic performance gains on PPBench and Sudoku-Extreme without retraining.

The paper 'Generative Recursive Reasoning' introduces a method that scales test-time compute by sampling multiple latent reasoning trajectories in parallel, enabling the model to explore diverse hypotheses and avoid deterministic collapse. This approach improves performance on tasks such as Sudoku, ARC AGI, N Queens, and graph coloring, and can also generate valid Sudoku boards and MNIST digits.

This paper demonstrates that transformers trained on Sudoku solving traces build structured world models organized by domain constraints, and identifies a sparse, monosemantic circuit responsible for the naked-single decision rule. The work provides a fully interpretable algorithmic account of transformer reasoning on a combinatorial task.

This article explores the mathematical underpinnings of Sudoku, explaining how the puzzle can be modeled as a graph theory vertex coloring problem. It details how algorithms like greedy search and backtracking are applied to solve these structures.