Abstract:A central task in knowledge compilation is to compile a CNF-SAT instance into a succinct representation format that allows efficient operations such as testing satisfiability, counting, or enumerating all solutions. Useful representation formats studied in this area range from ordered binary decision diagrams (OBDDs) to circuits in decomposable negation normal form (DNNFs). While it is known that there exist CNF formulas that require exponential size representations, the situation is less well studied for other types of constraints than Boolean disjunctive clauses. The constraint satisfaction problem (CSP) is a powerful framework that generalizes CNF-SAT by allowing arbitrary sets of constraints over any finite domain. The main goal of our work is to understand for which type of constraints (also called the constraint language) it is possible to efficiently compute representations of polynomial size. We answer this question completely and prove two tight characterizations of efficiently compilable constraint languages, depending on whether target format is structured. We first identify the combinatorial property of ``strong blockwise decomposability'' and show that if a constraint language has this property, we can compute DNNF representations of linear size. For all other constraint languages we construct families of CSP-instances that provably require DNNFs of exponential size. For a subclass of ``strong uniformly blockwise decomposable'' constraint languages we obtain a similar dichotomy for structured DNNFs. In fact, strong (uniform) blockwise decomposability even allows efficient compilation into multi-valued analogs of OBDDs and FBDDs, respectively. Thus, we get complete characterizations for all knowledge compilation classes between O(B)DDs and DNNFs.

New Advances in Compiling CNF into Decomposable Negation Normal Form

CCDD: A Tractable Representation for Model Counting and Uniform Sampling

A characterization of efficiently compilable constraint languages

Compiling Constraint Networks into Multivalued Decomposable Decision Graphs

Lower Bounds on Intermediate Results in Bottom-Up Knowledge Compilation

Translation Algorithm for Negative Literals in Conformant Planning.

The expressive language ALCNHR+K(D) for knowledge reasoning

CNF Encodings of Symmetric Functions

Compiling CP subproblems to MDDs and d-DNNFs

Structured d-DNNF Is Not Closed Under Negation

The splitting power of branching programs of bounded repetition and CNFs of bounded width

Extending Modern SAT Solvers for Enumerating All Models

Denotation-based Compositional Compiler Verification

ADDMC: Weighted Model Counting with Algebraic Decision Diagrams

The Power of Literal Equivalence in Model Counting

On Hashing-Based Approaches to Approximate DNF-Counting

Compiling Constraint Networks into AND/OR Multi-valued Decision Diagrams (AOMDDs)

On Tractable Representations of Binary Neural Networks

Pruning Boolean d-DNNF Circuits Through Tseitin-Awareness

From Total Assignment Enumeration to Modern SAT Solver

ConstraintFlow: A DSL for Specification and Verification of Neural Network Analyses