Learning SMT(LRA) Constraints using SMT Solvers

We introduce the problem of learning SMT(LRA) constraints from data. SMT(LRA) extends propositional logic with (in)equalities between numerical variables. Many relevant formal verification problems can be cast as SMT(LRA) instances and SMT(LRA) has supported recent developments in optimization and counting for hybrid Boolean and numerical domains. We introduce SMT(LRA) learning, the task of learning SMT(LRA) formulas from examples of feasible and infeasible instances, and we contribute INCAL, an exact non-greedy algorithm for this setting. Our approach encodes the learning task itself as an SMT(LRA) satisfiability problem that can be solved directly by SMT solvers. INCAL is an incremental algorithm that achieves exact learning by looking only at a small subset of the data, leading to significant speed-ups. We empirically evaluate our approach on both synthetic instances and benchmark problems taken from the SMT-LIB benchmarks repository.

[1]  James Cussens,et al.  Bayesian network learning by compiling to weighted MAX-SAT , 2008, UAI.

[2]  Kristin P. Bennett,et al.  Global Tree Optimization: A Non-greedy Decision Tree Algorithm , 2007 .

[3]  J. Ross Quinlan,et al.  Learning logical definitions from relations , 1990, Machine Learning.

[4]  Luc De Raedt,et al.  Inductive Logic Programming: Theory and Methods , 1994, J. Log. Program..

[5]  Silvia Tomasi,et al.  Optimization Modulo Theories with Linear Rational Costs , 2014, ACM Trans. Comput. Log..

[6]  Andrea Passerini,et al.  Structured learning modulo theories , 2014, Artif. Intell..

[7]  Jose M. Such,et al.  International Joint Conference on Artificial Intelligence (IJCAI) , 2016 .

[8]  Thomas G. Dietterich.  What is machine learning? , 2020, Archives of Disease in Childhood.

[9]  Andrew P. Witkin,et al.  Interactive physically-based manipulation of discrete/continuous models , 1995, SIGGRAPH.

[10]  Aaron Roth,et al.  Learning from Rational Behavior: Predicting Solutions to Unknown Linear Programs , 2016, NIPS.

[11]  Cesare Tinelli,et al.  Handbook of Satisfiability , 2021, Handbook of Satisfiability.

[12]  Nikolaj Bjørner,et al.  Z3: An Efficient SMT Solver , 2008, TACAS.

[13]  Michael I. Jordan,et al.  Advances in Neural Information Processing Systems 30 , 1995 .

[14]  P. Cochat,et al.  Et al , 2008, Archives de pediatrie : organe officiel de la Societe francaise de pediatrie.

[15]  Toby Walsh,et al.  Constraint Acquisition via Partial Queries , 2013, IJCAI.

[16]  Edward A. Lee.  Cyber Physical Systems: Design Challenges , 2008, 2008 11th IEEE International Symposium on Object and Component-Oriented Real-Time Distributed Computing (ISORC).

[17]  Guy Van den Broeck,et al.  Probabilistic Inference in Hybrid Domains by Weighted Model Integration , 2015, IJCAI.

[18]  Daniel Gooch,et al.  Communications of the ACM , 2011, XRDS.

[19]  Alberto Griggio,et al.  The MathSAT5 SMT Solver , 2013, TACAS.

[20]  Barry O'Sullivan,et al.  A SAT-Based Version Space Algorithm for Acquiring Constraint Satisfaction Problems , 2005, ECML.

[21]  Tommi S. Jaakkola,et al.  Learning Bayesian Network Structure using LP Relaxations , 2010, AISTATS.

[22]  Rajeev Alur,et al.  Syntax-guided synthesis , 2013, 2013 Formal Methods in Computer-Aided Design.

[23]  Nils J. Nilsson,et al.  Artificial Intelligence , 1974, IFIP Congress.

[24]  Roberto Battiti,et al.  Active Learning of Combinatorial Features for Interactive Optimization , 2011, LION.