Pranav Garg Research Statement
نویسنده
چکیده
My research agenda is to build verification technology that helps programmers write reliable, secure, and verified software. In particular, my research focuses on building automatic techniques that significantly lessen the burden on a programmer trying to prove her program secure or correct. The solutions I develop are learning based automatic software verification including machine learning algorithms for learning inductive program invariants, and reverse engineering a set of proof tactics from manual proofs to learn fully automatable natural proofs. My research impacts the building of verified software in the realms of software infrastructures and platforms that have many users, whose security and reliability is becoming increasingly important, and which include systems software such as operating systems, device drivers, mobile platforms, cloud infrastructures, and verification against specifications like race-freedom for parallel programs, memory safety and security.
منابع مشابه
Inferring Formal Properties of Production Key-Value Stores
Production distributed systems are challenging to formally verify, in particular when they are based on distributed protocols that are not rigorously described or fully understood. In this paper, we derive models and properties for two core distributed protocols used in eventually consistent production key-value stores such as Riak and Cassandra. We propose a novel modeling called certified pro...
متن کاملQuantified Data Automata on Skinny Trees: An Abstract Domain for Lists
We propose a new approach to heap analysis through an abstract domain of automata, called automatic shapes. The abstract do-domain of automata, called automatic shapes. The abstract domain uses a particular kind of automata, called quantified data automata on skinny trees (QSDAs), that allows to define universally quantified properties of singly-linked lists. To ensure convergence of the abstra...
متن کاملHorn-ICE Learning for Synthesizing Invariants and Contracts
We design learning algorithms for synthesizing invariants using Horn implication counterexamples (Horn-ICE), extending the ICE-learning model. In particular, we describe a decision-tree learning algorithm that learns from Horn-ICE samples, works in polynomial time, and uses statistical heuristics to learn small trees that satisfy the samples. Since most verification proofs can be modeled using ...
متن کاملLNCS 8559 - ICE: A Robust Framework for Learning Invariants
We introduce ICE, a robust learning paradigm for synthesizing invariants, that learns using examples, counter-examples, and implications, and show that it admits honest teachers and strongly convergent mechanisms for invariant synthesis. We observe that existing algorithms for black-box abstract interpretation can be interpreted as ICE-learning algorithms. We develop new strongly convergent ICE...
متن کاملICE: A Robust Framework for Learning Invariants
We introduce ICE, a robust learning paradigm for synthesizing invariants, that learns using examples, counter-examples, and implications, and show that it admits honest teachers and strongly convergent mechanisms for invariant synthesis. We observe that existing algorithms for black-box abstract interpretation can be interpreted as ICE-learning algorithms. We develop new strongly convergent ICE...
متن کامل