FaRAccel: FPGA-Accelerated Defense Architecture for Efficient Bit-Flip Attack Resilience in Transformer Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915585159331840 |
|---|---|
| author | Nazari, Najmeh Latibari, Banafsheh Saber Hosseini, Elahe Movafagh, Fatemeh Fang, Chongzhou Makrani, Hosein Mohammadi Gubbi, Kevin Immanuel Mahalanobis, Abhijit Rafatirad, Setareh Sayadi, Hossein Homayoun, Houman |
| author_facet | Nazari, Najmeh Latibari, Banafsheh Saber Hosseini, Elahe Movafagh, Fatemeh Fang, Chongzhou Makrani, Hosein Mohammadi Gubbi, Kevin Immanuel Mahalanobis, Abhijit Rafatirad, Setareh Sayadi, Hossein Homayoun, Houman |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_24985 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | FaRAccel: FPGA-Accelerated Defense Architecture for Efficient Bit-Flip Attack Resilience in Transformer Models Nazari, Najmeh Latibari, Banafsheh Saber Hosseini, Elahe Movafagh, Fatemeh Fang, Chongzhou Makrani, Hosein Mohammadi Gubbi, Kevin Immanuel Mahalanobis, Abhijit Rafatirad, Setareh Sayadi, Hossein Homayoun, Houman Cryptography and Security Artificial Intelligence 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. |
| title | FaRAccel: FPGA-Accelerated Defense Architecture for Efficient Bit-Flip Attack Resilience in Transformer Models |
| topic | Cryptography and Security Artificial Intelligence |
| url | https://arxiv.org/abs/2510.24985 |