Welcome to my homepage.
Welcome to my homepage.
I am a fourth-year graduate student at MIT advised by Martin Rinard (link). I am broadly interested in robust machine learning and program synthesis.
I also spent a year (2021-2022) concurrently at Google Brain as a Student Researcher, hosted by Sudip Roy (link) and Mangpo Phothilimthana (link) in the ML for ML Compilers group.
Previously, I was a Research Engineer at Reservoir Labs (which has since been acquired by Qualcomm AI Research). My research involved building compiler backends for task-based parallelism and heterogenous runtimes. Prior to that, I was an analyst at Weiss Asset Management (link), where I primarily worked in convertible bond arbitrage. (Please do NOT contact me about jobs in finance!)
I graduated summa cum laude from Yale University in 2016 with a B.S. in Mathematics and a combined B.S./M.S. in Computer Science, advised by Mariana Raykova (link) and Bryan Ford (link).
New preprint is up! Do language models acquire meaning from form? We contribute a formal perspective to this debate by leveraging techniques from PL. link
Charles Jin and Martin Rinard. "Evidence of Meaning in Language Models Trained on Programs." arXiv:2305.11169. 2023. link
Charles Jin, Zhang-Wei Hong, Farid Arthaud*, Idan Orzech*, and Martin Rinard. "Decentralized Inference via Capability Type sTructures in Cooperative Multi-Agent Systems." arXiv:2304.13957. 2023. link
Charles Jin, Zhang-Wei Hong, and Martin Rinard. "Toward Capability-Aware Cooperation for Decentralized Planning." Oral presentations at Decision Making in Multi-Agent Systems (DMMAS) and Workshop on Human Theory of Machines and Machine Theory of Mind for Human-Agent Teams (TOM4HAT) at IROS2022. Superceded by "Decentralized Inference via Capability Type Structures in Cooperative Multi-Agent Systems." link
Charles Jin and Martin Rinard. "Manifold Regularization for Locally Stable Deep Neural Networks." arXiv:2003.04286. 2020. link
Charles Jin, Melinda Sun, and Martin Rinard. "Incompatibility Clustering as a Defense Against Backdoor Poisoning Attacks." 11th International Conference on Learning Representations (ICLR 2023). link
Charles Jin, Phitchaya Mangpo Phothilimthana, and Sudip Roy, "αNAS: Neural Architecture Search using Property Guided Synthesis." Proceedings of the ACM on Programming Languages, Volume 6, Issue OOPSLA (OOPSLA 2022). link
Charles Jin and Martin Rinard. "Towards Context-Agnostic Learning Using Synthetic Data." Advances in Neural Information Processing Systems 34 (NeurIPS 2021). link
Limor Appelbaum, Alexandra Berg, Jose Cambronero, Thurston Dang, Charles Jin, Lori Zhang, Steven Kundrot, Matvey Palchuk, Laura Evans, Irving Kaplan, and Martin Rinard. "Development of a pancreatic cancer prediction model using a multinational medical records database." Journal of Clinical Oncology (JCO) 39:3_suppl, 394-394. 2021. link
Muthu Baskaran, Charles Jin, Benoit Meister, and Jonathan Springer. "Automatic Mapping and Optimization to Kokkos with Polyhedral Compilation." 2020 IEEE High Performance Extreme Computing Conference (HPEC20). Waltham, MA, USA. 2020. link
Charles Jin, Muthu Baskaran, Benoit Meister, and Jonathan Springer. "Automatic Parallelization to Asynchronous Task-Based Runtimes Through a Generic Runtime Layer." 2019 IEEE High Performance Extreme Computing Conference (HPEC19). Waltham, MA, USA. 2019. link
Charles Jin, Muthu Baskaran and Benoit Meister. "POSTER: Automatic Parallelization Targeting Asynchronous Task-Based Runtimes." 28th International Conference on Parallel Architectures and Compilation Techniques (PACT19). Seattle, WA, USA. 2019. link
Charles Jin and Muthu Baskaran. "Analysis of Explicit vs. Implicit Tasking in OpenMP Using Kripke." 2018 IEEE/ACM 4th International Workshop on Extreme Scale Programming Models and Middleware (ESPM2), held in conjunction with 2018 ACM/IEEE SuperComputing Conference (SC18). Dallas, TX, USA. 2018. link
"Automatic Code Generation to Dynamic Task-Based Runtimes: Recent Results." 10th Annual Concurrent Collections Workshop (CnC 2018).