Berkeley RDI hosted Agentic AI Summit, at UC Berkeley, with 2K+ in-person attendees and 40K+ joined online, the largest agentic AI gathering we know of, August 2025
Keynote at Microsoft Machine Learning, AI & Data Science Conference, June 2025. Towards building safe and secure AI - Lessons and Open Challenges
Keynote at ICLR, April 2025. Towards Building Safe and Secure AI: Lessons and Open Challenges
Invited talk at RSA, April 2025. Building Safe and Secure Agentic AI.
Keynote at Amazon Machine Learning Conference, October 2024. Towards Building Safe and Secure AI: Lessons and Open Challenges
Keynote at Graph the Planet, May 2024. Impact of Frontier AI on the Landscape of Cybersecurity
Invited talk at Stanford Workshop on the Governance of Open Foundation Models, February 2024. Impact of Frontier AI on the Landscape of Cybersecurity
Keynote at MLSys, Aug 2022. Towards Building a Responsible Data Economy
Keynote at IEEE Big Data, Dec 2021. Towards Building a Responsible Data Economy
Keynote at ACM CCS, Nov 2021. Towards Building a Responsible Data Economy
Keynote at ICDE, April 2021. Towards Building a Responsible Data Economy
Keynote at AAAI 2020, February 2020. AI and Security: Lessons, Challenges and Future Directions
Plenary keynote at KDD Deep Learning Day, August 2019. AI and Security: lessons, challenges and future directions
Keynote at CAV 2019, July 2019. Open Challenges in AI and Formal Verification
Invited talk at Deep Reinforcement Learning Summit, June, 2019. Secure Deep Reinforcement Learning
Keynote at ScaledML, March, 2019. AI and Security: Lessons and Future Directions
Invited talk at EmTech Digitals, March, 2019. AI and Security: Lessons and Future Directions
Distinguished lecture at Brown University Distinguished Lecture Series, October 2018. AI and Security: Lessons, Challenges and Future Directions
Keynote at IEEE Cybersecurity Development Conference, October 2018. Building and Deploying Secure Systems in Practice: Lessons, Challenges and Future Directions
Keynote at OReilly AI Conference, September, 2018. AI and security: Lessons, challenges, and future directions
Invited talk at Microsoft Research Faculty Summit, August, 2018. Towards secure, practical confidential computing with open source secure enclave
Keynote at ICML 2018, July, 2018. AI and Security: Lessons, Challenges and Future Directions
Keynote at Spark Summit, June, 2018. The Future of AI and Security
Keynote at ASPLOS workshop, March, 2018. Towards An Open-Source, Formally-Verified Secure Enclave
Invited talk at Representation Learning Workshop, March, 2017. Resilient Representation and Provable Generalization
Keynote at Apple Machine Learning Summit, February, 2017. Secure and Privacy-preserving data analytics and machine learning
Keynote at FMCAD, October, 2016. Formal Verification for Computer Security: Lessons Learned and Future Directions
Keynote at Qualcomm Research Day, June, 2015. Towards Security in a Connected World
Distinguished lecture at UPenn, April, 2015. No More Cat and Mouse: Towards Building Systems Secure by Construction
Job Openings
Positions are available for postdocs, research staff, staff programmers, and interns, in areas in security, privacy, and deep learning. If you are interested in applying for a position, please fill out the form here and then send an email to dawnsong.jobs@gmail.com if you are interested. Please do not send me emails directly without filling in the form.
For Berkeley undergrads, we enjoy having undergrads participate in our research projects to gain research experience. To be selected, normally we expect the students to have a GPA of at least 3.7 in EECS or math major or have extensive experience in areas related to the project. If interested, please apply through the URAP program. You are also welcome to fill in the form mentioned above at any time. Please do not send me emails directly without first applying through URAP and/or fill in the form above, unless instructed.
