We do research in the design and implementation of programming languages, program synthesis and repair, and computer-assisted formal reasoning about complex systems, at School of Computing at National University of Singapore.
Two postdoc positions are available in a joint project with Abhik Roychoudhury on automated program repair via static analysis and verification. Check out the project page for the details and get in touch!
Our current investigations follow the themes outlined below.
Program synthesis is an emerging research and technology paradigm for automatically deriving programs from user-provided declarative specifications, thereby significantly reducing the implementation effort required for producing correct-by-construction and efficient code.
Our recent work explored applications of state-of-the-art techniques for analysis, verification, and deductive proofs for fast and expressive program synthesis (check out the paper on SuSLik, and on the grand challenges in deductive synthesis of program with pointers) and for program repair. Our long-term agenda involves synthesis of correct concurrent and distributed programs by adopting our work on static analysis and logical foundations for reasoning about concurrent and distributed systems.
It is hard to overstate the significance and ubiquity of distributed services in many aspects of modern life, such as health care, online commerce, transportation, entertainment and cloud-based applications. Given the importance of distributed software and its complexity, it is vital in industry to have a rigorous verification methodology for establishing its correctness properties, ensuring that, once a distributed system is up and running, it will never go wrong and will eventually complete its goals.
Our recent work has established logical foundations for compositional verification of complex distributed protocols using a proof assistant. We have also produced the first mechanically verified proof of safety of Nakamoto consensus. Our ongoing work builds libraries and techniques for mechanised reasoning about probabilistic properties of distributed protocols and data structures employed by them. In particular, we have produced the first mechanised proof of the false-positive ratio for Bloom filters (see this blog post for more details).
In this line of research we apply core PL techniques, such as semantics, type systems, and abstract interpretation, for building safe and secure decentralised applications.
For instance, in our recent work, inspired by the verification ideas from Theme 2, we have developed a library for compositional construction of distributed protocols, allowing their modular testing and model-checking. By reflecting on the analogy between design principles of secure smart contracts (a particularly prominent class of decentralised applications) and concurrent software (also see the related ACSAC'18 and ISSTA'19 papers), in collaboration with industry partners we have developed Scilla, a functional smart contract language with strong safety guarantees. We have also developed a set of efficient compilation techniques for Scilla as well as a Coq-powered verification methodology for it.
Our ongoing research explores opportunities for (a) developing low-overhead abstractions for automated reasoning about distributed applications and (b) enhancing parallelism offered by modern distributed protocols via programming language techniques.
The paper on Certifying the Synthesis of Heap-Manipulating Programs will appear at ICFP'21.
Our invited position paper Deductive Synthesis of Programs with Pointers: Techniques, Challenges, Opportunities has appeared in the proceedings of CAV'21.
MComp Thesis. NUS School of Computing, 2022.
Capstone Thesis. Yale-NUS College, 2022.
Capstone Thesis. Yale-NUS College, 2022.
16th International Symposium on Functional and Logic Programming (FLOPS 2022). Taking place virtually, May 2022.
Principles of Blockchain Systems 2021. Pages 69–94. Morgan & Claypool. Author’s version.
26th ACM SIGPLAN International Conference on Functional Programming (ICFP 2021). Taking place virtually, August 2021.
2021 OCaml Users and Developers Workshop (OCaml 2021). Taking place virtually, August 2021.
33rd International Conference on Computer-Aided Verification (CAV 2021). Taking place virtually, July 2021.
2021 ACM SIGPLAN Conference on Programming Language Design and Implementation (PLDI 2021). Taking place virtually, June 2021.
2021 ACM SIGPLAN Conference on Programming Language Design and Implementation (PLDI 2021). Taking place virtually, June 2021. PLDI 2021 Distinguished Paper Award.