Research and ACM SRC Posters Archive

A Side-Channel Methodology for DL Model Recovery Using Power Traces


Poster Type: Research Posters

Author: Mohammed Sazzad Yousuf Sourab (Lamar University), Farah Ferdaus (Lamar University)

Abstract: Deep learning architecture represents proprietary intellectual property and serves as a basis for model extraction and adversarial attack techniques. By analyzing only the total GPU power consumption, sampled every 20ms via standard telemetry during inference, without access to weights, inputs, activations, or the computational graph, we can reconstruct the sequence of layer types executed. Our methodology involves segmenting the power trace, matching each segment against a library of per-layer power templates using k-Nearest Neighbors (k-NN), and refining the classification with a Hidden Markov Model (HMM) that encodes typical layer transition patterns with a Viterbi decoder. Evaluations across 21 models and 8-layer classes reveal an average layer-type sequence recovery accuracy of 60.1%, reaching up to 92% for some models.

Best Poster Finalist (BP): no
Poster: PDF
Poster Summary: PDF


Back to Poster Archive Listing