work ML for Earth Observation Temporal canopy height mapping at global scale using multi-sensor satellite data and vision transformers. Watermarks, Adversarial Defenses & Transferable Attacks A unified game-theoretic framework connecting watermarking, adversarial robustness, and transferable attacks via interactive protocols. Merlin-Arthur Classifiers Interpretable multi-agent classifiers with provable guarantees on feature relevance, inspired by interactive proof systems. Neural Concept Verifier Scaling Prover-Verifier Games to high-dimensional inputs via concept encodings for verifiable, interpretable AI. fun
