• Corpus ID: 18760406

Solving Sudoku Using Probabilistic Graphical Models

  title={Solving Sudoku Using Probabilistic Graphical Models},
  author={Sheehan Khan and Shahab Jabbari and Shahin Jabbari and Majid Ghanbarinejad},
Sudoku is a popular number puzzle. Here, we model the puzzle as a probabilistic graphical model and drive a modification to the well-known sum-product and max-product message passing to solve the puzzle. In addition, we propose a Sudoku solver utilizing a combination of message passing and Sinkhorn balancing and show that as Sudoku puzzles become larger, the impact of loopy propagation does not increase. 

Figures and Tables from this paper

Using Probabilistic Graphic Models to Solve NP-Complete Puzzle Problems
Using Probabilistic Graphical Models to Solve NP-complete Puzzle Problems Probabilistic Graphical Models (PGMs) are commonly used in machine learning to solve problems stemming from medicine,
Application of Message Passing and Sinkhorn Balancing Algorithms for Probabilistic Graphical Models
This project attempts to modify the sum-product message passing algorithm to solve Sudoku, and implements a solution using Sinkhorn balancing to overcome the impact of loopy propagation and compare its performance with the former.
An evolutionary approach to sudoku puzzles with filtered mutations
This paper proposes a GA with sophisticated genetic mutations that mitigate the worsening of fitness values in extremely difficult sudoku puzzles, in which only a few given numbers are provided.
Finding solutions to sudoku puzzles using human intuitive heuristics
  • N. Pillay
  • Computer Science
    South Afr. Comput. J.
  • 2012
The performance of the approach presented was found to be comparable to other methods used to solve Sudoku problems and in a number of cases produced better results.
Sudoku at the Intersection of Classical and Quantum Computing
Even when quantum computers are widely available, it is unlikely that we will simply use them to solve all our problems. Rather we would use each computer for the types of problems it is best suited
Rešavanje problema CSP tehnikama svodjenja na problem SAT
The main aim of this thesis is to improve CSP solving techniques that are using reduction to SAT, and two new hybrid encodings of CSPs to SAT are presented and they combine good sides of the existingencodings.
A hardware generator for factor graph applications
  • James Demma, P. Athanas
  • Computer Science
    2014 International Conference on ReConFigurable Computing and FPGAs (ReConFig14)
  • 2014
A framework is presented for rapid prototyping of hardware implementations of FG-based applications, and the framework has been applied to construct a Sudoku solver.
Recurrent Relational Networks for Complex Relational Reasoning
R recurrent relational networks are introduced which increase the suite of solvable tasks to those that require an order of magnitude more steps of relational reasoning and are applied to the BaBi textual QA dataset solving 19/20 tasks.
Recurrent Relational Networks
The recurrent relational network is introduced, a general purpose module that operates on a graph representation of objects that can augment any neural network model with the capacity to do many-step relational reasoning.
End-to-end information extraction from business documents
A recurrent neural network model that can capture long range context and compare it to a baseline logistic regression model corresponding to the current CloudScan production system are described.


Solving Sudoku Using Combined Message Passing Algorithms
In this paper we apply message-passing algorithms to solve Sudoku puzzles. We provide explicit expression for the sum-product algorithm and the max-product algorithm and analyze the difference
Multiple Constraint Satisfaction by Belief Propagation: An Example Using Sudoku
  • T. Moon, J. Gunther
  • Mathematics
    2006 IEEE Mountain Workshop on Adaptive and Learning Systems
  • 2006
The popular Sudoku puzzle bears structural resemblance to the problem of decoding linear error correction codes: solution is over a discrete set, and several constraints apply. We express the
Passing Messages to Lonely Numbers
  • H. Bauke
  • Computer Science
    Computing in Science & Engineering
  • 2008
Sudoku, a popular number puzzle, is a simple optimization problem that message-passing algorithms can help solve and is an ideal vehicle to demonstrate these methods' strengths and limitations.
Complexity and Completeness of Finding Another Solution and Its Application to Puzzles
  • T. Yato, Takahiro Seta
  • Computer Science, Mathematics
    IEICE Trans. Fundam. Electron. Commun. Comput. Sci.
  • 2003
This paper considers n-ASP, the problem to find another solution when n solutions are given, and considers ASP-completeness, the completeness with respect to the parsimonious reductions which allow polynomial-time transformation.
Solving, rating and generating Sudoku puzzles with GA
The results presented in this paper seem to support the conclusion that the objectives of this study are reasonably well met with genetic algorithm optimization.
On the optimality of solutions of the max-product belief-propagation algorithm in arbitrary graphs
It is shown that the assignment based on a fixed point of max-product is a "neighborhood maximum" of the posterior probabilities: the posterior probability of the max- product assignment is guaranteed to be greater than all other assignments in a particular large region around that assignment.
Generating Satisfiable Problem Instances
This paper proposes a generator that only outputs satisfiable problem instances and shows how one can finely control the hardness of the satisfiable instances by establishing a connection between problem hardness and a new kind of phase transition phenomenon in the space of problem instances.
Iterative Decoding for Sudoku and Latin Square Codes
This paper presents a new class of q-ary erasure-correcting codes based on Latin and Sudoku squares of order q, and an iterative decoding algorithm similar to the one used for the Low Density Parity
Building Industrial Applications with Constraint Programming
An overview of real-life applications developed with constraint logic programming and the concepts of global constraints, meta-heuristics and constraint visualisation, which are central to large scale constraint solving are introduced.
Graph Theory with Applications
The burgeoning of Graph Theory was first aware when I studied the 1940 paper of Brooks, Smith, Stone and Tutte in the Duke Mathematical Journal, ostensibly on squared rectangles, all in the Quest of the Perfect Square.