Research

Inferential Learning & Privacy

As artificial intelligence algorithms make rapid strides in deriving seemingly hidden inferences from outwardly harmless data, the privacy of individuals is often compromised, not through hacking or identity theft but through learning and inference. Legal protections for privacy only extend to information that is stolen, not that which is inferred. Our research develops data processing and decision making mechanisms backed by statistical inference theory that balance optimally the level of desired privacy and the loss in utility in the data or decision making processes.

Ruochi Zhang and Parv Venkitasubramaniam. “Optimal distribution mapping for inference privacy.” 2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC). IEEE, 2018. pdf

Ruochi Zhang and Parv Venkitasubramaniam. “Mutual-Information-Private Online Gradient Descent Algorithm.” 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2018. pdf

Ruochi Zhang, and Parv Venkitasubramaniam. “Optimal Multi-Source Inference Privacy—A Generalized Lloyd-Max Algorithm.” 2018 56th Annual Allerton Conference on Communication, Control, and Computing (Allerton). IEEE, 2018. pdf 

Omid Javidbakht and Parv Venkitasubramaniam, “Inference Resistant Policy Design for Markov Decision Processes, “ 2018 IEEE Global Signal and Image Processing Conference (GlobalSIP). IEEE, 2018. 

Venkitasubramaniam, Parv, Jiyun Yao, and Parth Pradhan. “Information-theoretic security in stochastic control systems.” Proceedings of the IEEE 103.10 (2015): 1914-1931. pdf

Venkitasubramaniam, Parv. “Privacy in stochastic control: A markov decision process perspective.” 2013 51st Annual Allerton Conference on Communication, Control, and Computing (Allerton). IEEE, 2013 pdf

Venkitasubramaniam, Parv. “Decision making under privacy restrictions.” 52nd IEEE Conference on Decision and Control. IEEE, 2013. pdf

Pradhan, Parth, and Parv Venkitasubramaniam. “Under the radar attacks in dynamical systems: Adversarial privacy utility tradeoffs.” 2014 IEEE Information Theory Workshop (ITW 2014). IEEE, 2014. pdf

 

Leave a Reply

Powered by: Wordpress
Skip to toolbar