Publication Date
Spring 2025
Degree Type
Master's Project
Degree Name
Master of Science in Computer Science (MSCS)
Department
Computer Science
First Advisor
Saptarshi Sengupta
Second Advisor
Sayma Akther
Third Advisor
Mohammad Masum
Keywords
Time Series Forecasting, Quantization, Adversarial Attacks, Model Compression, Robustness, Edge Deployment.
Abstract
This work explores securing and optimization of Transformer-based time series forecasting models. We employ several quantization techniques, including quantization-aware training (QAT), and tested the robustness of quantized models by adversarially attacking them. The preliminary results of this work, in our controlled setup, indicate that quantized models outperform the full precision model in terms of robustness against adversarial attacks. They achieved this robustness while showing a very minimal decrease in their forecasting performance.
Recommended Citation
Siddiqui, Fahad, "Comparative Analysis of Adversarial Permeability in CPU-native, QAT and ONNX-based Quantized Transformer Models" (2025). Master's Projects. 1562.
DOI: https://doi.org/10.31979/etd.ffn9-m4gs
https://scholarworks.sjsu.edu/etd_projects/1562
