Search

Home
Publications
CV
News

Light Dark Automatic

Vitaly Shmatikov

Latest

Deep-Research Agents Can Be Poisoned via User-Generated Content
Adversarial Hubness in Multi-Modal Retrieval
How to Steal Reasoning Without Reasoning Traces
Adversarial Decoding: Generating Readable Documents for Adversarial Objectives
Self-interpreting Adversarial Images
Adversarial Illusions in Multi-Modal Embeddings

© 2026 Me. This work is licensed under CC BY NC ND 4.0

Published with Wowchemy — the free, open source website builder that empowers creators.

Cite