Martin Rinard

Control Jujutsu

2015

Control flow integrity (CFI) has been proposed as an approach to defend against control-hijacking memory corruption attacks. CFI works by assigning tags to indirect branch targets statically and checking them at runtime. Coarse-grained enforcements of …

Characterizing Developer Use of Automatically Generated Patches

2019

We present a study that characterizes the way developers use automatically generated patches when fixing software defects. Our study tasked two groups of developers with repairing defects in C programs. Both groups were provided with …

Manifold Regularization for Locally Stable Deep Neural Networks

2020 · arXiv (Cornell University)

We apply concepts from manifold regularization to develop new regularization techniques for training locally stable deep neural networks. Our regularizers are based on a sparsification of the graph Laplacian which holds with high probability when …

Program Synthesis Guided Reinforcement Learning

2021 · arXiv (Cornell University)

A key challenge for reinforcement learning is solving long-horizon planning and control problems. Recent work has proposed leveraging programs to help guide the learning algorithm in these settings. However, these approaches impose a high manual …

Incremental Computation for Efficient Programmable Inference in Probabilistic Programs

2025 · Proceedings of the ACM on Programming Languages

Inference in probabilistic programs generally requires evaluating many possible program executions to find those of high posterior density. To scale inference to large datasets, it is crucial that expensive intermediate results are shared across these …