Computer Science

The Stevens School of Computing is pleased to announce that Assistant Professor Michael Greenberg has received an award from the National Science Foundation to support his research on improving the reliability of widely used computer programs.

The project, titled “Collaborative Research: FMitF: Track I: Reasoning About Shell Scripts and Their Effects in Context,” focuses on shell scripts — small programs that automate important computing tasks behind the scenes in systems used by developers, scientists and system administrators worldwide. While these scripts are widely used, they can be difficult to test thoroughly and errors may only appear when a program is already running. In some cases, these errors can lead to lost work, corrupted files or unexpected changes to computing environments.

Greenberg’s research aims to develop methods that can analyze shell scripts before they are executed, helping to identify potential issues in advance. By better understanding how scripts interact with files and system tools, the project seeks to improve the safety and reliability of everyday computing workflows. The work will also produce open-source tools and benchmarks to support broader use and evaluation by the research community.

This award supports ongoing work in building more reliable and dependable computing systems.

Assistant Professor Tegan Brennan has received a National Science Foundation (NSF) CAREER Award of $574,990 for her project, “SaTC: Privacy-Aware Just-In-Time Compilation.”

This research addresses a key challenge in software security: ensuring that the compiler optimizations used to improve performance do not unintentionally introduce side-channel vulnerabilities that could expose sensitive information through indirect signals such as program execution time.

Professor Brennan’s project will create a privacy-aware just-in-time compilation framework that balances security with performance. The framework will identify optimizations that may introduce risk, quantify the strength of potential vulnerabilities, and apply data-driven strategies to decide when optimizations should be applied or reversed.

By integrating privacy awareness into the compilation process, this work will help developers focus on writing correct and efficient software without having to manage complex low-level security risks. The project will advance both software performance and privacy protection, contributing to more secure computing for end users.

Professor Zhaozhuo Xu has received two National Science Foundation (NSF) awards supporting cutting-edge research in machine learning and artificial intelligence infrastructure.

Xu’s project, “CRII: III: Compression-Aware Computing for Sustainable Machine Learning Model Inference” ($171,387), aims to improve the efficiency of machine learning model inference by developing a compression-aware computing framework. As AI models continue to grow in size and computational demands, this research will create self-aware models that can detect, adapt to, and recover from lossy compression — techniques like sparsification and quantization that allow large models to run on consumer devices. The project’s goal is to address the trade-off between model accuracy and efficiency, enabling cost-effective, energy-efficient AI systems.

His second NSF project, “Collaborative Research: Planning: CIRC: Planning-C: Building Foundation Model Infrastructure for Feature Interaction Discovery” ($49,991), focuses on building shared AI infrastructure for biological research. This work seeks to connect computer science and biology communities by developing scalable foundation model infrastructure designed for discovering feature interactions — the ways different biological features combine to influence outcomes. The project includes workshops and collaborative research to advance foundation model training, datasets, and benchmarks that support data-driven discovery in complex biological systems.

The Amazon Research Awards program has provided funding to a groundbreaking project led by Ping Wang, an assistant professor in Stevens’ Computer Science program. The project, titled "Leveraging Large Language Models for Reasoning Augmented Searching on Domain-Specific NoSQL Database," aims to advance natural language querying (NLQ) techniques for NoSQL databases.

The research focuses on using modern deep learning and semantic parsing techniques to translate natural language questions into structured query language (SQL) queries. Despite the prevalence of efficient NoSQL databases, there has been limited development of NLQ tools for these systems. Wang's team will conduct a comprehensive analysis of NoSQL databases and integrate recent advances in multi-hop reasoning on knowledge graphs and large language models (LLMs) to propose a reasoning-augmented searching framework for ElasticSearch.

This innovative project promises to enhance NLQ techniques on NoSQL databases, offering significant benefits for various real-world applications. The funding from Amazon Research will support the team's efforts to bridge the gap in NLQ tool development and exploit the potential of NoSQL databases.