Securing AI Agents with Information-Flow Control
Manuel Costa, Boris Köpf, Aashish Kolluri, Andrew Paverd, Mark Russinovich, Ahmed Salem, Shruti Tople, Lukas Wutschitz, Santiago Zanella-Béguelin
May 2025
Lukas Wutschitz
Principal Researcher
Papiers de Collaboration
-
-
-
The Canary’s Echo: Auditing Privacy Risks of LLM-Generated Synthetic Text
Matthieu Meeus, Lukas Wutschitz, Santiago Zanella-Béguelin, Reza Shokri, Shruti Tople
February 2025
-
The Canary’s Echo: Auditing Privacy Risks of LLM-Generated Synthetic Text
Matthieu Meeus, Lukas Wutschitz, Santiago Zanella-Béguelin, Shruti Tople, Reza Shokri
ICML 2025 | February 2025
-
-
-
Permissive Information-Flow Analysis for Large Language Models
Shoaib Ahmed Siddiqui, Radhika Gaonkar, Boris Köpf, David Krueger, Andrew Paverd, Ahmed Salem, Shruti Tople, Lukas Wutschitz, Menglin Xia, Santiago Zanella-Béguelin
October 2024
-
Closed-Form Bounds for DP-SGD against Record-level Inference
Giovanni Cherubin, Boris Köpf, Andrew Paverd, Shruti Tople, Lukas Wutschitz, Santiago Zanella-Béguelin
USENIX Security Symposium | August 2024
-
-
-
Bayesian Estimation of Differential Privacy
Santiago Zanella-Béguelin, Lukas Wutschitz, Shruti Tople, Ahmed Salem, Victor Ruehle, Andrew Paverd, Mohammad Naseri, Boris Köpf, Daniel Jones
2023 International Conference on Machine Learning | July 2023
Rédacteur(s) en chef: Barbara Engelhardt, Emma Brunskill, Kyunghyun Cho
-
Analyzing Leakage of Personally Identifiable Information in Language Models
Nils Lukas, Ahmed Salem, Robert Sim, Shruti Tople, Lukas Wutschitz, Santiago Zanella-Béguelin
2023 IEEE Symposium on Security and Privacy | May 2023
-
-
-
Differentially Private Model Compression
Fatemehsadat Mireshghallah, Arturs Backurs, Huseyin Inan, Lukas Wutschitz, Janardhan (Jana) Kulkarni
NeurIPS 2022 | November 2022
-
dp-transformers: Training transformer models with differential privacy
Huseyin Inan, Andre Manoel, Lukas Wutschitz
July 2022
-
Differentially private fine-tuning of language models
Da Yu, Saurabh Naik, Arturs Backurs, Sivakanth Gopi, Huseyin Inan, Gautam Kamath, Janardhan (Jana) Kulkarni, Yin Tat Lee, Andre Manoel, Lukas Wutschitz, Sergey Yekhanin, Huishuai Zhang
ICLR 2022 | April 2022
-
-
-
Numerical Composition of Differential Privacy
Sivakanth Gopi, Yin Tat Lee, Lukas Wutschitz
NeurIPS 2021 | June 2021
Spotlight at NeurIPS 2021.
-
Training Data Leakage Analysis in Language Models
Huseyin A. Inan, Osman Ramadan, Lukas Wutschitz, Daniel Jones, Victor Rühle, James Withers, Robert Sim
February 2021
-
-
-
Analyzing Information Leakage of Updates to Natural Language Models
Santiago Zanella-Béguelin, Lukas Wutschitz, Shruti Tople, Victor Ruehle, Andrew Paverd, Olga Ohrimenko, Boris Köpf, Marc Brockschmidt
ACM Conference on Computer and Communication Security (CCS) | November 2020
-
-
-
Securing AI Agents with Information-Flow Control
Manuel Costa, Boris Köpf, Aashish Kolluri, Andrew Paverd, Mark Russinovich, Ahmed Salem, Shruti Tople, Lukas Wutschitz, Santiago Zanella-Béguelin
May 2025
-
The Canary’s Echo: Auditing Privacy Risks of LLM-Generated Synthetic Text
Matthieu Meeus, Lukas Wutschitz, Santiago Zanella-Béguelin, Reza Shokri, Shruti Tople
February 2025
-
The Canary’s Echo: Auditing Privacy Risks of LLM-Generated Synthetic Text
Matthieu Meeus, Lukas Wutschitz, Santiago Zanella-Béguelin, Shruti Tople, Reza Shokri
ICML 2025 | February 2025
-
Permissive Information-Flow Analysis for Large Language Models
Shoaib Ahmed Siddiqui, Radhika Gaonkar, Boris Köpf, David Krueger, Andrew Paverd, Ahmed Salem, Shruti Tople, Lukas Wutschitz, Menglin Xia, Santiago Zanella-Béguelin
October 2024
-
Closed-Form Bounds for DP-SGD against Record-level Inference
Giovanni Cherubin, Boris Köpf, Andrew Paverd, Shruti Tople, Lukas Wutschitz, Santiago Zanella-Béguelin
USENIX Security Symposium | August 2024
-
Bayesian Estimation of Differential Privacy
Santiago Zanella-Béguelin, Lukas Wutschitz, Shruti Tople, Ahmed Salem, Victor Ruehle, Andrew Paverd, Mohammad Naseri, Boris Köpf, Daniel Jones
2023 International Conference on Machine Learning | July 2023
Rédacteur(s) en chef: Barbara Engelhardt, Emma Brunskill, Kyunghyun Cho
-
Analyzing Leakage of Personally Identifiable Information in Language Models
Nils Lukas, Ahmed Salem, Robert Sim, Shruti Tople, Lukas Wutschitz, Santiago Zanella-Béguelin
2023 IEEE Symposium on Security and Privacy | May 2023
-
Differentially Private Model Compression
Fatemehsadat Mireshghallah, Arturs Backurs, Huseyin Inan, Lukas Wutschitz, Janardhan (Jana) Kulkarni
NeurIPS 2022 | November 2022
-
dp-transformers: Training transformer models with differential privacy
Huseyin Inan, Andre Manoel, Lukas Wutschitz
July 2022
-
Training Data Leakage Analysis in Language Models
Huseyin A. Inan, Osman Ramadan, Lukas Wutschitz, Daniel Jones, Victor Rühle, James Withers, Robert Sim
February 2021
-
Analyzing Information Leakage of Updates to Natural Language Models
Santiago Zanella-Béguelin, Lukas Wutschitz, Shruti Tople, Victor Ruehle, Andrew Paverd, Olga Ohrimenko, Boris Köpf, Marc Brockschmidt
ACM Conference on Computer and Communication Security (CCS) | November 2020
-
-
-
Securing AI Agents with Information-Flow Control
Manuel Costa, Boris Köpf, Aashish Kolluri, Andrew Paverd, Mark Russinovich, Ahmed Salem, Shruti Tople, Lukas Wutschitz, Santiago Zanella-Béguelin
May 2025
-
The Canary’s Echo: Auditing Privacy Risks of LLM-Generated Synthetic Text
Matthieu Meeus, Lukas Wutschitz, Santiago Zanella-Béguelin, Reza Shokri, Shruti Tople
February 2025
-
Permissive Information-Flow Analysis for Large Language Models
Shoaib Ahmed Siddiqui, Radhika Gaonkar, Boris Köpf, David Krueger, Andrew Paverd, Ahmed Salem, Shruti Tople, Lukas Wutschitz, Menglin Xia, Santiago Zanella-Béguelin
October 2024
-
Closed-Form Bounds for DP-SGD against Record-level Inference
Giovanni Cherubin, Boris Köpf, Andrew Paverd, Shruti Tople, Lukas Wutschitz, Santiago Zanella-Béguelin
USENIX Security Symposium | August 2024
-
Bayesian Estimation of Differential Privacy
Santiago Zanella-Béguelin, Lukas Wutschitz, Shruti Tople, Ahmed Salem, Victor Ruehle, Andrew Paverd, Mohammad Naseri, Boris Köpf, Daniel Jones
2023 International Conference on Machine Learning | July 2023
Rédacteur(s) en chef: Barbara Engelhardt, Emma Brunskill, Kyunghyun Cho
-
Analyzing Leakage of Personally Identifiable Information in Language Models
Nils Lukas, Ahmed Salem, Robert Sim, Shruti Tople, Lukas Wutschitz, Santiago Zanella-Béguelin
2023 IEEE Symposium on Security and Privacy | May 2023
-
dp-transformers: Training transformer models with differential privacy
Huseyin Inan, Andre Manoel, Lukas Wutschitz
July 2022
-
Numerical Composition of Differential Privacy
Sivakanth Gopi, Yin Tat Lee, Lukas Wutschitz
NeurIPS 2021 | June 2021
Spotlight at NeurIPS 2021.
-
Training Data Leakage Analysis in Language Models
Huseyin A. Inan, Osman Ramadan, Lukas Wutschitz, Daniel Jones, Victor Rühle, James Withers, Robert Sim
February 2021
-
Analyzing Information Leakage of Updates to Natural Language Models
Santiago Zanella-Béguelin, Lukas Wutschitz, Shruti Tople, Victor Ruehle, Andrew Paverd, Olga Ohrimenko, Boris Köpf, Marc Brockschmidt
ACM Conference on Computer and Communication Security (CCS) | November 2020
-
-
-
Differentially Private Model Compression
Fatemehsadat Mireshghallah, Arturs Backurs, Huseyin Inan, Lukas Wutschitz, Janardhan (Jana) Kulkarni
NeurIPS 2022 | November 2022
-
Differentially private fine-tuning of language models
Da Yu, Saurabh Naik, Arturs Backurs, Sivakanth Gopi, Huseyin Inan, Gautam Kamath, Janardhan (Jana) Kulkarni, Yin Tat Lee, Andre Manoel, Lukas Wutschitz, Sergey Yekhanin, Huishuai Zhang
ICLR 2022 | April 2022
-
Numerical Composition of Differential Privacy
Sivakanth Gopi, Yin Tat Lee, Lukas Wutschitz
NeurIPS 2021 | June 2021
Spotlight at NeurIPS 2021.
-
-
-
Training Data Leakage Analysis in Language Models
Huseyin A. Inan, Osman Ramadan, Lukas Wutschitz, Daniel Jones, Victor Rühle, James Withers, Robert Sim
February 2021
-
-
-
Numerical Composition of Differential Privacy
Sivakanth Gopi, Yin Tat Lee, Lukas Wutschitz
NeurIPS 2021 | June 2021
Spotlight at NeurIPS 2021.
-
-
-
Securing AI Agents with Information-Flow Control
Manuel Costa, Boris Köpf, Aashish Kolluri, Andrew Paverd, Mark Russinovich, Ahmed Salem, Shruti Tople, Lukas Wutschitz, Santiago Zanella-Béguelin
May 2025
-
Permissive Information-Flow Analysis for Large Language Models
Shoaib Ahmed Siddiqui, Radhika Gaonkar, Boris Köpf, David Krueger, Andrew Paverd, Ahmed Salem, Shruti Tople, Lukas Wutschitz, Menglin Xia, Santiago Zanella-Béguelin
October 2024
-
-
-
The Canary’s Echo: Auditing Privacy Risks of LLM-Generated Synthetic Text
Matthieu Meeus, Lukas Wutschitz, Santiago Zanella-Béguelin, Shruti Tople, Reza Shokri
ICML 2025 | February 2025
-
Closed-Form Bounds for DP-SGD against Record-level Inference
Giovanni Cherubin, Boris Köpf, Andrew Paverd, Shruti Tople, Lukas Wutschitz, Santiago Zanella-Béguelin
USENIX Security Symposium | August 2024
-
Bayesian Estimation of Differential Privacy
Santiago Zanella-Béguelin, Lukas Wutschitz, Shruti Tople, Ahmed Salem, Victor Ruehle, Andrew Paverd, Mohammad Naseri, Boris Köpf, Daniel Jones
2023 International Conference on Machine Learning | July 2023
Rédacteur(s) en chef: Barbara Engelhardt, Emma Brunskill, Kyunghyun Cho
-
Analyzing Leakage of Personally Identifiable Information in Language Models
Nils Lukas, Ahmed Salem, Robert Sim, Shruti Tople, Lukas Wutschitz, Santiago Zanella-Béguelin
2023 IEEE Symposium on Security and Privacy | May 2023
-
Differentially Private Model Compression
Fatemehsadat Mireshghallah, Arturs Backurs, Huseyin Inan, Lukas Wutschitz, Janardhan (Jana) Kulkarni
NeurIPS 2022 | November 2022
-
Differentially private fine-tuning of language models
Da Yu, Saurabh Naik, Arturs Backurs, Sivakanth Gopi, Huseyin Inan, Gautam Kamath, Janardhan (Jana) Kulkarni, Yin Tat Lee, Andre Manoel, Lukas Wutschitz, Sergey Yekhanin, Huishuai Zhang
ICLR 2022 | April 2022
-
Numerical Composition of Differential Privacy
Sivakanth Gopi, Yin Tat Lee, Lukas Wutschitz
NeurIPS 2021 | June 2021
Spotlight at NeurIPS 2021.
-
Analyzing Information Leakage of Updates to Natural Language Models
Santiago Zanella-Béguelin, Lukas Wutschitz, Shruti Tople, Victor Ruehle, Andrew Paverd, Olga Ohrimenko, Boris Köpf, Marc Brockschmidt
ACM Conference on Computer and Communication Security (CCS) | November 2020
-
-
-
The Canary’s Echo: Auditing Privacy Risks of LLM-Generated Synthetic Text
Matthieu Meeus, Lukas Wutschitz, Santiago Zanella-Béguelin, Reza Shokri, Shruti Tople
February 2025
-
dp-transformers: Training transformer models with differential privacy
Huseyin Inan, Andre Manoel, Lukas Wutschitz
July 2022
-
Training Data Leakage Analysis in Language Models
Huseyin A. Inan, Osman Ramadan, Lukas Wutschitz, Daniel Jones, Victor Rühle, James Withers, Robert Sim
February 2021
-