Browsing by Author "Vardi, Moshe"

Now showing 1 - 5 of 5
  • Thumbnail Image
    Item
    Drug Target Optimization in Chronic Myeloid Leukemia Using Innovative Computational Platform
    (Nature Publishing Group, 2015) Chuang, Ryan; Hall, Benjamin A.; Benque, David; Cook, Byron; Ishtiaq, Samin; Piterman, Nir; Taylor, Alex; Vardi, Moshe; Koschmieder, Steffen; Gottgens, Berthold; Fisher, Jasmin
    Chronic Myeloid Leukemia (CML) represents a paradigm for the wider cancer field. Despite the fact that tyrosine kinase inhibitors have established targeted molecular therapy in CML, patients often face the risk of developing drug resistance, caused by mutations and/or activation of alternative cellular pathways. To optimize drug development, one needs to systematically test all possible combinations of drug targets within the genetic network that regulates the disease. The BioModelAnalyzer (BMA) is a user-friendly computational tool that allows us to do exactly that. We used BMA to build a CML network-model composed of 54 nodes linked by 104 interactions that encapsulates experimental data collected from 160 publications. While previous studies were limited by their focus on a single pathway or cellular process, our executable model allowed us to probe dynamic interactions between multiple pathways and cellular outcomes, suggest new combinatorial therapeutic targets, and highlight previously unexplored sensitivities to Interleukin-3.
  • Thumbnail Image
    Item
    Intro to Logic
    (Rice University, 2010-06-11) Barland, Ian; Kolaitis, Phokion; Vardi, Moshe; Felleisen, Matthias; Greiner, John
    An introduction to reasoning with propositional and first-order logic, with applications to computer science. Part of the TeachLogic Project (www.teachlogic.org).
  • Thumbnail Image
    Item
    Model Checking Concurrent Programs
    (Rice University, 2007-04-12) Barland, Ian; Vardi, Moshe; Greiner, John
    Introduction to the use of Promela and SPIN as a model-checker for verifying concurrent programs. The underlying transition system is used as a model for the interpretation of Linear Temporal Logic (LTL) formulas.
  • Thumbnail Image
    Item
    Quantitative Reasoning on Hybrid Formulas with Dynamic Programming
    (2022-08-12) Phan, Vu Hoang Nguyen; Vardi, Moshe
    We perform quantitative reasoning via Stochastic Satisfiability (SSAT), a problem that involves the existential quantifier, the random quantifier, and Boolean formulas. SSAT generalizes the Weighted Satisfiability (WSAT), Weighted Model Counting (WMC), Weighted Projected Model Counting (WPMC), and Existential-Random Satisfiability (ERSAT) problems. SSAT has been used to model power-transmission reliability, machine-learning fairness, and other applications. Prior work on SSAT mainly focuses on Conjunctive Normal Form (CNF). A CNF formula is a conjunctively interpreted set of disjunctive clauses. We study a more general format, XOR-extended Conjunctive Normal Form (XOR-CNF). An XOR-CNF formula is a set of hybrid constraints that include XOR clauses and disjunctive clauses. Native support for XOR is desirable because this constraint type is prevalent in several applications, such as cryptography and circuit design. Although XOR-CNF can be converted into CNF by adding linearly many auxiliary variables and disjunctive clauses, this apparently small overhead actually decimates the performance of CNF solvers in our empirical evaluation. Our solvers natively handle hybrid constraints with algebraic decision diagrams, which provide compact representations of XOR and disjunctive clauses. To solve a big set of constraints, we employ dynamic programming. This approach partitions the constraint set and solves resultant subsets before merging partial solutions into a final answer. Given an XOR-CNF formula, we obtain a requested solution using a Project-Join Tree (PJT) that specifies how to conjoin clauses and project out variables of the formula. Existentially quantified variables are maximally projected, and randomly quantified variables are summatively projected. Our framework operates in two phases. First, the planning phase constructs a PJT from an XOR-CNF formula. Second, the execution phase traverses the constructed PJT to compute a solution. This framework is versatile because different planners and executors can be combined. Also, we may use a single plan with multiple executions: a one-layer PJT with WSAT and WMC, or a two-layer PJT with WPMC and ERSAT. We empirically demonstrate that our exact solvers meaningfully contribute to the state of the art. On 1606 one-layer instances from inference and planning applications, adding our framework leads to significant reductions in mean PAR-2 scores of solver portfolios: 6.0% for WSAT and 27.7% for WMC. On 613 two-layer instances from sampling and counting applications, the score reductions are 5.8% for WPMC and 10.7% for ERSAT. Future work includes merging our SSAT framework with another PJT-based framework that solves minimum-maximum satisfiability. We will also consider adding more layers of quantification to support a more general problem, maximum model counting.
  • Thumbnail Image
    Item
    Weighted Model Counting with Algebraic Decision Diagrams
    (2019-12-04) Phan, Vu Hoang Nguyen; Vardi, Moshe
    We present an algorithm to compute exact literal-weighted model counts of Boolean formulas in conjunctive normal form. Our algorithm employs dynamic programming and uses algebraic decision diagrams as the primary data structure. We implement this technique in ADDMC, a new model counter. We empirically evaluate various heuristics that can be used with ADDMC. We then compare ADDMC to state-of-the-art exact weighted model counters (Cachet, c2d, d4, and miniC2D) on 1914 standard model counting benchmarks and show that ADDMC significantly improves the virtual best solver.