Project Grant 2442825

Award Date 7/15/25
Completion Date 6/30/30
Dollars Obligated $379K
Federal Grant Program
47.070
Assistance Type
Project Grant
Place of Performance
Penn State University, PA 16802, USA
Similar Awards
This NSF CISE (CFDA 47.070) $600,000 Project Grant awarded to The Pennsylvania State University, doing business as Penn State, will develop methods to reduce privacy risks while enabling users to benefit from high-quality AI-based services. The key goals are to create mathematically-guaranteed privacy-preserving encodings or embeddings that can be used as input for AI-based applications, such as artistic image generation, medical diagnosis, and assistive technologies for the visually impaired....
This Project Grant from the National Science Foundation's Computer and Information Science and Engineering program, CFDA #47.070, provides $265,054 to Weill Medical College of Cornell University to develop a consolidated framework for computational privacy and machine learning from October 1, 2022 to September 30, 2026. The framework aims to comprehensively consider optimal tradeoffs between privacy protections and critical machine learning properties like predictive utility, fairness, and...
This Project Grant award of $241,682 from the National Science Foundation's (NSF) Computer and Information Science and Engineering (CISE) program supports the development of a consolidated framework for computational privacy and machine learning. The project aims to comprehensively address the trade-offs between privacy protection, utility, fairness, and distributed learning in real-world machine learning applications, with a focus on medical research. The award will fund the research team at...
This National Science Foundation (NSF) Computer and Information Science and Engineering (CISE) Federal Grant Award (CFDA 47.070) provides $592,294 to The Pennsylvania State University to develop secure, robust, and end-user driven prediction-aware counterfactual explanations (CFEs) for machine learning (ML) models. The 3-year project aims to address key limitations of current CFE techniques, including the potential for intellectual property theft, inability to handle model updates, and lack of...
This Project Grant award from the National Science Foundation (NSF) under the Computer and Information Science and Engineering (CISE) program (CFDA 47.070) provides $279,959 to Carnegie Mellon University to advance the frontiers of differential privacy algorithms for private learning and synthetic data generation. The 5-year research project aims to develop a theoretical framework to better capture practical privacy scenarios, design practical privacy-preserving algorithms, and create auditing...
The National Science Foundation (NSF) awarded a $423,204 Project Grant under the Computer and Information Science and Engineering (CISE) program to The Pennsylvania State University (Penn State) for a 4-year collaborative research project titled "COLLABORATIVE RESEARCH: SATC: CORE: MEDIUM: DIFFERENTIALLY PRIVATE SQL WITH FLEXIBLE PRIVACY MODELING, MACHINE-CHECKED SYSTEM DESIGN, AND ACCURACY OPTIMIZATION." The goal is to develop an open-source, customizable system for preserving privacy...
This National Science Foundation (NSF) Computer and Information Science and Engineering (CISE) Program (CFDA 47.070) project grant, awarded to North Carolina State University, is focused on enhancing privacy in federated learning, which is an AI approach enabling knowledge sharing without compromising data privacy. The $220,258 grant, awarded on October 1, 2024, aims to address vulnerabilities in federated learning schemes that may leak sensitive information through improper privacy...
This Project Grant award of $600,000 from the National Science Foundation's (NSF) Computer and Information Science and Engineering (CISE) program supports the development of a security-focused framework to protect collaborative scientific computing in machine learning as a service (MLaaS) environments. The key research thrusts of the project include: Robust model protection techniques to hinder reverse engineering of machine learning models while preserving their utility. Behavioral monitoring...
The National Science Foundation (NSF) awarded a $581,966 Project Grant under the Computer and Information Science and Engineering (CFDA 47.070) program to the University of Pittsburgh. The grant, awarded on May 1, 2024, with a completion date of April 30, 2027, supports the development of a hardware-software co-design framework to address performance and memory space issues for privacy protection in cloud-based deep learning recommendation systems (DLRMs). Key project objectives include: 1)...
The Pennsylvania State University received a $100,000 Project Grant award from the National Science Foundation Division of Social and Economic Science under the Social, Behavioral, and Economic Sciences federal grant program (CFDA 47.075). The award will support research from June 2023 to May 2026 to address concerns about privacy risks from personal information sharing on social media, especially for low-income and underrepresented groups. The university will study how socioeconomic status...

This Project Grant award from the National Science Foundation's (NSF) Computer and Information Science and Engineering (CISE) program, with a CFDA number of 47.070, provides $379,224 to The Pennsylvania State University (Penn State) from July 2025 to June 2030. The project aims to develop better methods for assessing privacy risks in machine learning (ML) models trained on tabular data, which are commonly used in privacy-sensitive domains like healthcare and finance. The key products and services to be delivered include:

  1. A framework to systematically audit attribute inference vulnerabilities in ML models by introducing an adaptable adversary model, designing novel attack algorithms, and developing an automated ML privacy auditing tool.

  2. A mathematical formalization to characterize disparity in attribute inference risks, along with novel attack techniques that exploit this disparity and target more vulnerable groups.

  3. Robust defenses for different phases of the ML pipeline, including data pre-processing, training, and inference, to mitigate attribute inference attacks and disparity in both centralized and federated learning settings.

The project aims to increase the privacy of people whose data is used in ML models, enabling their safer use in important applications. No sub-awards are planned for this grant.

Generated 8/5/25, 3:16 AM