The Tsetlin Machine - A Game Theoretic Bandit Driven Approach to Optimal Pattern Recognition with Propositional Logic

Although simple individually, artificial neurons provide state-of-the-art performance when interconnected in deep networks. Arguably, the Tsetlin Automaton is an even simpler and more versatile learning mechanism, capable of solving the multi-armed bandit problem. Merely by means of a single integer as memory, it learns the optimal action in stochastic environments through increment and decrement operations. In this paper, we introduce the Tsetlin Machine, which solves complex pattern recognition problems with propositional formulas, composed by a collective of Tsetlin Automata. To eliminate the longstanding problem of vanishing signal-to-noise ratio, the Tsetlin Machine orchestrates the automata using a novel game. Further, both inputs, patterns, and outputs are expressed as bits, while recognition and learning rely on bit manipulation, simplifying computation. Our theoretical analysis establishes that the Nash equilibria of the game align with the propositional formulas that provide optimal pattern recognition accuracy. This translates to learning without local optima, only global ones. In five benchmarks, the Tsetlin Machine provides competitive accuracy compared with SVMs, Decision Trees, Random Forests, Naive Bayes Classifier, Logistic Regression, and Neural Networks. We further demonstrate how the propositional formulas facilitate interpretation. We believe the combination of high accuracy, interpretability, and computational simplicity makes the Tsetlin Machine a promising tool for a wide range of domains.

[1]  Shigeo Abe DrEng.  Pattern Classification , 2001, Springer London.

[2]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[3]  B. John Oommen,et al.  Deterministic Learning Automata Solutions to the Equipartitioning Problem , 1988, IEEE Trans. Computers.

[4]  Ole-Christoffer Granmo,et al.  The Convolutional Tsetlin Machine , 2019, ArXiv.

[5]  M. Thathachar,et al.  Networks of Learning Automata: Techniques for Online Stochastic Optimization , 2003 .

[6]  P. Anandan,et al.  Pattern-recognizing stochastic learning automata , 1985, IEEE Transactions on Systems, Man, and Cybernetics.

[7]  David G. Stork,et al.  Pattern Classification (2nd ed.) , 1999 .

[8]  Ole-Christoffer Granmo,et al.  Combining finite learning automata with GSAT for the satisfiability problem , 2010, Eng. Appl. Artif. Intell..

[9]  H. Robbins.  Some aspects of the sequential design of experiments , 1952 .

[10]  Yann LeCun,et al.  Regularization of Neural Networks using DropConnect , 2013, ICML.

[11]  Ole-Christoffer Granmo,et al.  Stochastic Learning for SAT- Encoded Graph Coloring Problems , 2010, Int. J. Appl. Metaheuristic Comput..

[12]  Ole-Christoffer Granmo,et al.  Solving two-armed Bernoulli bandit problems using a Bayesian learning automaton , 2010, Int. J. Intell. Comput. Cybern..

[13]  Jürgen Schmidhuber,et al.  Deep learning in neural networks: An overview , 2014, Neural Networks.

[14]  B. John Oommen,et al.  Stochastic searching on the line and its applications to parameter learning in nonlinear optimization , 1997, IEEE Trans. Syst. Man Cybern. Part B.

[15]  Cynthia Rudin,et al.  A Bayesian Framework for Learning Rule Sets for Interpretable Classification , 2017, J. Mach. Learn. Res..

[16]  Mohammad Reza Meybodi,et al.  A streaming sampling algorithm for social activity networks using fixed structure learning automata , 2018, Applied Intelligence.

[17]  Leonard Kleinrock,et al.  Using Finite State Automata to Produce Self-Optimization and Self-Control , 1996, IEEE Trans. Parallel Distributed Syst..

[18]  B. John Oommen,et al.  A Solution to the Stochastic Point Location Problem in Metalevel Nonstationary Environments , 2008, IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics).

[19]  B. John Oommen,et al.  A Novel Strategy for Solving the Stochastic Point Location Problem Using a Hierarchical Searching Scheme , 2014, IEEE Transactions on Cybernetics.

[20]  J. Gittins.  Bandit processes and dynamic allocation indices , 1979 .

[21]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[22]  Peter Auer,et al.  Using Confidence Bounds for Exploitation-Exploration Trade-offs , 2003, J. Mach. Learn. Res..

[23]  Sudip Misra,et al.  Routing Bandwidth-Guaranteed Paths in MPLS Traffic Engineering: A Multiple Race Track Learning Approach , 2007, IEEE Transactions on Computers.

[24]  Yoshua Bengio,et al.  Gradient-based learning applied to document recognition , 1998, Proc. IEEE.

[25]  Anis Yazidi,et al.  Learning-Automaton-Based Online Discovery and Tracking of Spatiotemporal Event Patterns , 2013, IEEE Transactions on Cybernetics.

[26]  Darrell D. E. Long,et al.  Theory of finite automata with an introduction to formal languages , 1989 .

[27]  Lihong Li,et al.  An Empirical Evaluation of Thompson Sampling , 2011, NIPS.

[28]  Ole-Christoffer Granmo,et al.  Solving the Satisfiability Problem Using Finite Learning Automata , 2007, Int. J. Comput. Sci. Appl..

[29]  Anis Yazidi,et al.  Service selection in stochastic environments: a learning-automaton based solution , 2012, Applied Intelligence.

[30]  B. John Oommen,et al.  Learning Automata-Based Solutions to the Nonlinear Fractional Knapsack Problem With Applications to Optimal Resource Allocation , 2007, IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics).

[31]  B. Oommen,et al.  On the analysis of a random walk-jump chain with tree-based transitions and its applications to faulty dichotomous search , 2018 .

[32]  J. Ross Quinlan,et al.  Induction of Decision Trees , 1986, Machine Learning.

[33]  David S. Leslie,et al.  Optimistic Bayesian Sampling in Contextual-Bandit Problems , 2012, J. Mach. Learn. Res..

[34]  Kumpati S. Narendra,et al.  Learning automata - an introduction , 1989 .

[35]  B. John Oommen,et al.  Solving Stochastic Nonlinear Resource Allocation Problems Using a Hierarchy of Twofold Resource Allocation Automata , 2010, IEEE Transactions on Computers.

[36]  Stephen A. Cook,et al.  The complexity of theorem-proving procedures , 1971, STOC.

[37]  B. John Oommen,et al.  Optimal sampling for estimation with constrained resources using a learning automaton-based solution for the nonlinear fractional knapsack problem , 2010, Applied Intelligence.

[38]  Kush R. Varshney,et al.  Interpretable Two-level Boolean Rule Learning for Classification , 2015, ArXiv.

[39]  MengChu Zhou,et al.  Symmetrical Hierarchical Stochastic Searching on the Line in Informative and Deceptive Environments , 2017, IEEE Transactions on Cybernetics.

[40]  W. R. Thompson.  ON THE LIKELIHOOD THAT ONE UNKNOWN PROBABILITY EXCEEDS ANOTHER IN VIEW OF THE EVIDENCE OF TWO SAMPLES , 1933 .

[41]  Ole-Christoffer Granmo,et al.  A Two-Armed Bandit Collective for Hierarchical Examplar Based Mining of Frequent Itemsets with Applications to Intrusion Detection , 2014, Trans. Comput. Collect. Intell..

[42]  Yee Whye Teh,et al.  A Fast Learning Algorithm for Deep Belief Nets , 2006, Neural Computation.