Light-Bound Transformers: Hardware-Anchored Robustness for Silicon-Photonic Computer Vision Systems
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866913007161835520 |
|---|---|
| author | Chen, Xuming Najafi, Deniz Zhou, Chengwei Mercati, Pietro Roohi, Arman Imani, Mohsen Nikdast, Mahdi Angizi, Shaahin Datta, Gourav |
| author_facet | Chen, Xuming Najafi, Deniz Zhou, Chengwei Mercati, Pietro Roohi, Arman Imani, Mohsen Nikdast, Mahdi Angizi, Shaahin Datta, Gourav |
| contents | Deploying Vision Transformers (ViTs) on near-sensor analog accelerators demands training pipelines that are explicitly aligned with device-level noise and energy constraints. We introduce a compact framework for silicon-photonic execution of ViTs that integrates measured hardware noise, robust attention training, and an energy-aware processing flow. We first characterize bank-level noise in microring-resonator (MR) arrays, including fabrication variation, thermal drift, and amplitude noise, and convert these measurements into closed-form, activation-dependent variance proxies for attention logits and feed-forward activations. Using these proxies, we develop Chance-Constrained Training (CCT), which enforces variance-normalized logit margins to bound attention rank flips, and a noise-aware LayerNorm that stabilizes feature statistics without changing the optical schedule. These components yield a practical ``measure $\rightarrow$ model $\rightarrow$ train $\rightarrow$ run'' pipeline that optimizes accuracy under noise while respecting system energy limits. Hardware-in-the-loop experiments with MR photonic banks show that our approach restores near-clean accuracy under realistic noise budgets, with no in-situ learning or additional optical MACs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_04330 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Light-Bound Transformers: Hardware-Anchored Robustness for Silicon-Photonic Computer Vision Systems Chen, Xuming Najafi, Deniz Zhou, Chengwei Mercati, Pietro Roohi, Arman Imani, Mohsen Nikdast, Mahdi Angizi, Shaahin Datta, Gourav Emerging Technologies Deploying Vision Transformers (ViTs) on near-sensor analog accelerators demands training pipelines that are explicitly aligned with device-level noise and energy constraints. We introduce a compact framework for silicon-photonic execution of ViTs that integrates measured hardware noise, robust attention training, and an energy-aware processing flow. We first characterize bank-level noise in microring-resonator (MR) arrays, including fabrication variation, thermal drift, and amplitude noise, and convert these measurements into closed-form, activation-dependent variance proxies for attention logits and feed-forward activations. Using these proxies, we develop Chance-Constrained Training (CCT), which enforces variance-normalized logit margins to bound attention rank flips, and a noise-aware LayerNorm that stabilizes feature statistics without changing the optical schedule. These components yield a practical ``measure $\rightarrow$ model $\rightarrow$ train $\rightarrow$ run'' pipeline that optimizes accuracy under noise while respecting system energy limits. Hardware-in-the-loop experiments with MR photonic banks show that our approach restores near-clean accuracy under realistic noise budgets, with no in-situ learning or additional optical MACs. |
| title | Light-Bound Transformers: Hardware-Anchored Robustness for Silicon-Photonic Computer Vision Systems |
| topic | Emerging Technologies |
| url | https://arxiv.org/abs/2604.04330 |