publications

2026

1. 

   Security in the Age of AI Teammates: An Empirical Study of Agentic Pull Requests on GitHub

   Mohammed Latif Siddiq, Xinye Zhao, Vinicius Carvalho Lopes, and 2 more authors

   2026

2. 

   An Empirical Study on Remote Code Execution in Machine Learning Model Hosting Ecosystems

   Mohammed Latif Siddiq, Tanzim Hossain Romel, Natalie Sekerak, and 2 more authors

   2026

2025

1. 

   Quantum-Based SMT Solving for String Theory

   Beatrice Casey, Joanna C. S. Santos, and Andrew Hennessee

   HPDC ’25: Proceedings of the 34th International Symposium on High-Performance Parallel and Distributed Computing, Sep 2025

   DOI

2. 

   A Survey of Source Code Representations for Machine Learning-Based Cybersecurity Tasks

   Beatrice Casey, Joanna C. S. Santos, and George Perry

   ACM Comput. Surv., Mar 2025

   DOI

3. 

   An Empirical Study of Safetensors’ Usage Trends and Developers’ Perceptions

   Beatrice Casey, Kaia Damian, Andrew Cotaj, and 1 more author

   Mar 2025

2024

1. 

   A Large-Scale Exploit Instrumentation Study of AI/ML Supply Chain Attacks in Hugging Face Models

   Beatrice Casey, Joanna C. S. Santos, and Mehdi Mirakhorli

   Mar 2024

2. 

   Franc: A Lightweight Framework for High-Quality Code Generation

   Mohammed Latif Siddiq, Beatrice Casey, and Joanna C. S. Santos

   In 2024 IEEE International Conference on Source Code Analysis and Manipulation (SCAM), Mar 2024

   DOI

2022

1. 

   PiSpy: An affordable, accessible, and flexible imaging platform for the automated observation of organismal biology and behavior

   Benjamin I. Morris, Marcy J. Kittredge, Bea Casey, and 5 more authors

   PLOS ONE, Oct 2022

   Abs DOI

   A great deal of understanding can be gleaned from direct observation of organismal growth, development, and behavior. However, direct observation can be time consuming and influence the organism through unintentional stimuli. Additionally, video capturing equipment can often be prohibitively expensive, difficult to modify to one’s specific needs, and may come with unnecessary features. Here, we describe PiSpy, a low-cost, automated video acquisition platform that uses a Raspberry Pi computer and camera to record video or images at specified time intervals or when externally triggered. All settings and controls, such as programmable light cycling, are accessible to users with no programming experience through an easy-to-use graphical user interface. Importantly, the entire PiSpy system can be assembled for less than $100 using laser-cut and 3D-printed components. We demonstrate the broad applications and flexibility of PiSpy across a range of model and non-model organisms. Designs, instructions, and code can be accessed through an online repository, where a global community of PiSpy users can also contribute their own unique customizations and help grow the community of open-source research solutions.