5 papers
One stout to rule them all: Reconciling artificial intelligence, data science and malted alcoholic beverages
Dmitrii Usynin, Elena Shmakova, Michael Rheinberger
Beer is a phenomenal beverage. It has previously shaped the history of many peoples, states and cultures. The beauty of beer is its versatility. Starting from the original implemen…
A LINDDUN-based Privacy Threat Modeling Framework for GenAI
Qianying Liao, Jonah Bellemans, Laurens Sion +6
As generative AI (GenAI) systems become increasingly prevalent across various technological stacks, the question of how such systems handle sensitive and personal data flows become…
PII Jailbreaking in LLMs via Activation Steering Reveals Personal Information Leakage
Krishna Kanth Nakka, Xue Jiang, Dmitrii Usynin +1
This paper investigates privacy jailbreaking in LLMs via steering, focusing on whether manipulating activations can bypass LLM alignment and alter response behaviors to privacy rel…
Mitigating Backdoor Attacks using Activation-Guided Model Editing
Felix Hsieh, Huy H. Nguyen, AprilPyone MaungMaung +2
Backdoor attacks compromise the integrity and reliability of machine learning models by embedding a hidden trigger during the training process, which can later be activated to caus…
Naturally Private Recommendations with Determinantal Point Processes
Jack Fitzsimons, Agustín Freitas Pasqualini, Robert Pisarczyk +1
Often we consider machine learning models or statistical analysis methods which we endeavour to alter, by introducing a randomized mechanism, to make the model conform to a differe…