Bailey 🕸️🕷️🦇

95 Followers
162 Following
117 Posts

Research in designing human centered privacy preserving data science, e.g., MPC, PSI, privacy for ML, etc. Assistant Professor at University of Alberta

bkacsmar.github.io

I recently spoke with Adam White @AmiiThinks about my research, and being one of Amii's new fellows. Check it out below if you want to hear about privacy and machine learning!

https://www.amii.ca/latest-from-amii/meet-fellows-bailey-kacsmar/

Meet the Fellows: Bailey Kacsmar

Learn more about the research and work of Bailey Kacsmar, one of the latest Fellows to join Amii’s team of world-class researchers. Hear Bailey talk about the importance of human-centred privacy in machine learning, and tackling the "privacy paradox."

Alberta Machine Intelligence Institute
The 2023 Best HotPETs talk is:
Bridging the Gap between Privacy Incidents and PETs Shannon Veitch (ETH Zurich), Lena Csomor (ETH Zurich), Alexander Viand (ETH Zurich), Anwar Hithnawi(ETH Zurich), and Bailey Kacsmar (University of Alberta) #hotpets23
The 2023 runner up for Best HotPETs talk is: Female mHealth Apps: Opportunities and Challenges The authors are Ina Kaleva (King’s College London), Lisa Mekioussa Malki (King’s College London), Dilisha Patel (Global Disability Innovation Hub, University College London), Mark Warner (University College London), and Ruba Abu-Salma (King’s College London) #hotpets23
Heard a senior academic recently say, in regards to not being able to afford talent for intern positions, that "Grad students these days have different standards. When I was in grad school, I had roommates and ate ramen". To be crystal clear: renting a room in a house *with 5 roommates* in Boston costs $1,000/month, not including utilities. $15/hr before taxes means 70+ hours of work *just to pay rent*.

The runner-up for the 2023 Caspar Bowden PET Award is: “Extracting Training Data from Large Language Models” #pets23

Congratulations to the authors: Nicholas Carlini, Florian Tramèr, Eric Wallace, Matthew Jagielski, Ariel Herbert-Voss, Katherine Lee, Adam Roberts, Tom Brown, Dawn Song, Úlfar Erlingsson, Alina Oprea, Colin Raffel (USENIX Security ’21)

The 2023 Caspar Bowden PET Award winner is: "The 2020 Census Disclosure Avoidance System TopDown Algorithm"

Congratulations to the authors: John M. Abowd, Robert Ashmead, Ryan Cumings-Menon, Simson Garfinkel, Micah Heineck, Christine Heiss, Robert Johns, Daniel Kifer, Philip Leclerc, Ashwin Machanavajjhala, Brett Moran, William Sexton, Matthew Spence, Pavel Zhuravlev (Harvard Data Science Review '22)

#pets23

The 2023 best artifact is "HeLayers: A Tile Tensors Framework for Large Neural Networks on Encrypted Data" by Ehud Aharoni, Allon Adir, Moran Baruch, Nir Drucker, Gilad Ezov, Ariel Farkash, Lev Greenberg, Ramy Masalha, Guy Moshkowich, Dov Murik, Hayim Shaul, Omri Soceanu. #pets23

#pets23 The runners-up for the Andreas Pfitzmann Best Student Paper Award are:

"Convolutions in Overdrive: Maliciously Secure Convolutions for MPC" by Marc Rivinius, Pascal Reisert, Sebastian Hasler, Ralf Küsters. #pets23

"Not Your Average App: A Large-scale Privacy Analysis of Android Browsers" by Amogh Pradeep, Álvaro Feal, Julien Gamba, Ashwin Rao, Martina Lindorfer, Narseo Vallina-Rodriguez, David Choffnes. #pets23

"Story Beyond the Eye: Glyph Positions Break PDF Text Redaction" by Maxwell Bland, Anushya Iyer, Kirill Levchenko. #pets23

#pets23 The winners of the Andreas Pfitzmann Best Student Paper Award are:

"Researchers’ Experiences in Analyzing Privacy Policies: Challenges and Opportunities" by Abraham Mhaidli, Selin Fidan, An Doan, Gina Herakovic, Mukund Srinath, Lee Matheson, Shomir Wilson, and Florian Schaub. #pets23

"Lox: Protecting the Social Graph in Bridge Distribution" by Lindsey Tulloch and Ian Goldberg. #pets23

"Everybody's Looking for SSOmething: A large-scale evaluation on the privacy of OAuth authentication on the web" by Yana Dimova, Tom Van Goethem, and Wouter Joosen. #pets23

The "ethics in software engineering" classes I've seen use things like Therac-25 as examples, but it seems pretty esoteric to most students, and with near-daily mass shootings in the news, the loss of life does not impress. Does anyone currently teach a software ethics course that focuses primarily on memetic threats, e.g., developers' responsibility/liability w.r.t. disinformation/radicalization that leads to mass killings or targeted harassment? If so, pointers welcome.