Abstract:Complex-Valued Neural Networks (CVNNs) have significant advantages in handling tasks that involve complex numbers. However, existing CVNNs are unable to quantify predictive uncertainty. We propose, for the first time, dropout-based Bayesian Complex-Valued Neural Networks (BayesCVNNs) to enable uncertainty quantification for complex-valued applications, exhibiting broad applicability and efficiency for hardware implementation due to modularity. Furthermore, as the dual-part nature of complex values significantly broadens the design space and enables novel configurations based on layer-mixing and part-mixing, we introduce an automated search approach to effectively identify optimal configurations for both real and imaginary components. To facilitate deployment, we present a framework that generates customized FPGA-based accelerators for BayesCVNNs, leveraging a set of optimized building blocks. Experiments demonstrate the best configuration can be effectively found via the automated search, attaining higher performance with lower hardware costs compared with manually crafted models. The optimized accelerators achieve approximately 4.5x and 13x speedups on different models with less than 10% power consumption compared to GPU implementations, and outperform existing work in both algorithm and hardware aspects. Our code is publicly available at: this https URL.
| Comments: | Accepted to 63rd ACM/IEEE Design Automation Conference (DAC '26). 7 pages, 6 figures |
| Subjects: | Hardware Architecture (cs.AR); Machine Learning (cs.LG) |
| Cite as: | arXiv:2604.19993 [cs.AR] |
| (or arXiv:2604.19993v1 [cs.AR] for this version) | |
| https://doi.org/10.48550/arXiv.2604.19993 arXiv-issued DOI via DataCite (pending registration) |
|
| Journal reference: | 63rd ACM/IEEE Design Automation Conference (DAC '26), July 2026 |
| Related DOI: | https://doi.org/10.1145/3770743.3804197
DOI(s) linking to related resources |
From: Zehuan Zhang [view email]
[v1]
Tue, 21 Apr 2026 21:06:00 UTC (4,398 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。