Publications
2025
-
MoEcho: Exploiting Side-Channel Attacks to Compromise User Privacy in Mixture-of-Experts LLMsCCS 2025
-
Graph in the Vault: Protecting Edge GNN Inference with Trusted Execution EnvironmentDAC 2025
-
MACPruning: Dynamic Operation Pruning to Mitigate Side-Channel DNN Model ExtractionHOST 2025 🎯 Best Paper Nomination
-
Probe-Me-Not: Protecting Pre-trained Encoders from Malicious ProbingNDSS 2025
2024
-
GraphCroc: Cross-Correlation Autoencoder for Graph Structural ReconstructionNeurIPS 2024
-
Non-transferable Pruning for Controlled Model ReuseECCV 2024
2023
-
VertexSerum: Poisoning Graph Neural Networks for Link InferenceICCV 2023
-
EMShepherd: Detecting Adversarial Samples via Side-channel LeakageASIACCS 2023 🏆 Distinguished Paper Award
2022
-
A cross-platform cache timing attack framework via deep learningDATE 2022 🎯 Best Paper Nomination
-
Ran$Net: An Anti-Ransomware Methodology based on Cache Monitoring and Deep LearningGLSVLSI 2022
2021
-
Neural Architecture Search for Adversarial Robustness via Learnable PruningFrontiers in High Performance Computing
-
Spatio-temporal point processes with attention for traffic congestion event modelingIEEE Trans. on Intelligent Transportation Systems
-
Deep Fourier Kernel for Self-Attentive Point ProcessesAISTATS 2021
2018
-
Persistent fault analysis on block ciphersIACR Trans. on Cryptographic Hardware and Embedded Systems