Team Toronto around Nicolas Papernot and Adam Dziedzic won the ICLR’22 Spotlight Award for “Increasing the Cost of Model Extraction with Calibrated Proof of Work”.
ICRL’22 Spotlight Award Read More »
The University of Southern California Team around Salman Avestimehr and Murali Annavaram won the FL-AAAI ’22 Best Paper Award for “SSFL: Tackling Label Deficiency in Federated Learning via Personalized Self-Supervision“.
Best Paper Award at the FL-AAAI ’22 Read More »
On November 4th there was a Private AI Roundtable with the following pitch talks:
Phillip Rieger, Technical University of Darmstadt
Mitigating Backdoor Attacks in Federated Learning Through Deep Model Inspection
Florian Kerschbaum, University of Waterloo
Is Differential Privacy what you want to protect privacy in ML?
Adam Dziedzic, University of Toronto
CaPC: from proof-of-concept to real-world applications
Slides here
Mojan Javaheripi, UC San Diego
Defending against fault injection attacks on DNNs.
Saurav Prakash, USC
How to detect adversaries in FL without breaching privacy?
Private AI Institute Roundtable Read More »
In this episode of Cyber Security Inside by Intel Business, Camille Morhardt gets into the details with Ahmad-Reza Sadeghi, who is a professor at TU Darmstadt in Germany, and JV Rajendran, an assistant professor at Texas A&M. A fascinating discussion that you should definitely check out.
Cyber Security Inside – Video cast with Ahmad-Reza Sadeghi Read More »
Prof. Ahmad-Reza Sadeghi and his team have been honored with the Intel Academic Leadership Award at the prestigious USENIX Security conference 2021.
The award recognizes the significant impact on shaping international cyber security research, particularly in hardware-based security, and transferring knowledge and insights to academic and industrial security community.
In addition, Intel also recognizes with the award the work of TU Darmstadt together with Texas A&M University for designing and organizing the world’s largest hardware security competition HACK@Event. The Award was announced and presented at the internationally renowned cyber security conference USENIX Security 2021.
The U.S. semiconductor giant Intel thus honors the excellent research work and the international impact of Prof. Sadeghi’s work. “We are very proud of this award”, says a delighted Ahmad-Reza Sadeghi. “Throughout my career I have worked with many large multinational enterprises. My encounters with Intel were always accompanied with passion for deep and visionary research. Intel seeks to work with top innovative minds worldwide, no matter where they are” says the computer science professor in laudation of the award.
Intel Academic Leadership Award 2021 Read More »
The annual international conference for AI, machine learning, and cybersecurity experts from academia and industry hosted by Avast, The Private AI Collaborative Research Institute and the Czech Technical University in Prague AI Center FEE CTU.
The international conference focus on AI for online privacy protection and security. As technology evolves, new opportunities, challenges, and ethical dilemmas constantly present themselves to those working and researching at the intersection of AI and cybersecurity.
Find more information on the CyberSec&AI Connected website.
You can register here.
CyberSec&AI Connected 2021 Read More »
Reza Shokri, National Universitiy of Singapore received the prestigious IEEE Security and Privacy Test of Time Award 2021 for his work “Quantifying Location Privacy”. He also received the VMWare early career faculty award 2021.
He presented the “CaPC Learning: Confidential and Private Collaborative Learning” protocol that allows private consultation at the FLOW seminar. Here you will find the talk.
CaPC Learning: Confidential and Private Collaborative Learning Read More »
Salman Avestimehr provided a series of invited talks on “Trustworthy and Scalable Federated Learning” at UC Berkeley, FLOW Seminar series. Please find the talk here.
Trustworthy and Scalable Federated Learning Read More »
Salman Avestimehr and his group (USC) has released FedNLP, FedGraphNN, and FedCV research platforms in FedML for FL research in NLP and GraphNN, and CV application domains. They include a diverse set of advanced tasks in each application domain, and motivate several key research problems that arise in each setting.
Release of research platforms Read More »
