FaRAccel: FPGA-Accelerated Defense Architecture for Efficient Bit-Flip Attack Resilience in Transformer Models
Authors:
- Najmeh NazariNNNajmeh Nazari
- Banafsheh Saber LatibariBLBanafsheh Saber Latibari
- Elahe HosseiniEHElahe Hosseini
- Fatemeh MovafaghFMFatemeh Movafagh
- Chongzhou FangCFChongzhou Fang
- Hosein Mohammadi MakraniHMHosein Mohammadi Makrani
- Kevin Immanuel GubbiKGKevin Immanuel Gubbi
- Abhijit MahalanobisAMAbhijit Mahalanobis
- Setareh RafatiradSRSetareh Rafatirad
- Hossein SayadiHossein Sayadi
- Houman HomayounHHHouman Homayoun
Preprint
Abstract
Forget and Rewire (FaR) methodology has demonstrated strong resilience against Bit-Flip Attacks (BFAs) on Transformer-based models by obfuscating critical parameters through dynamic rewiring of linear layers. However, the application of FaR introduces non-negligible performance and memory overheads, primarily due to the runtime modification of activation pathways and the lack of hardware-level optimization. To overcome these limitations, we propose FaRAccel, a novel hardware accelerator architecture implemented on FPGA, specifically designed to offload and optimize FaR operations. FaRAccel integrates reconfigurable logic for dynamic activation rerouting, and lightweight storage of rewiring configurations, enabling low-latency inference with minimal energy overhead. We evaluate FaRAccel across a suite of Transformer models and demonstrate substantial reductions in FaR inference latency and improvement in energy efficiency, while maintaining the robustness gains of the original FaR methodology. To the best of our knowledge, this is the first hardware-accelerated defense against BFAs in Transformers, effectively bridging the gap between algorithmic resilience and efficient deployment on real-world AI platforms.Keywords
All Related Versions
This publication has 6 references indexed in Scilit:
- MLB-MAC: Multi-Level Binary MAC Array for Energy Efficient ML AcceleratorsIEEE Transactions on Circuits and Systems I: Regular Papers, 2025
- Large Language Models for Opioid-Induced Respiratory Depression Prediction in Hospitalized Patients: A Retrospective StudyACM Transactions on Computing for Healthcare, 2025
- Optimizing Vision Transformers: Unveiling ‘Focus and Forget’ for Enhanced Computational EfficiencyIEEE Access, 2025
- OpenGeMM: A Highly-Efficient GeMM Accelerator Generator with Lightweight RISC-V Control and Tight Memory Coupling2025
- FaCT-LSTM: Fast and Compact Ternary Architecture for LSTM Recurrent Neural NetworksIEEE Design & Test, 2021
- RowHammer: A RetrospectiveIEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 2019