Multi-Dimensional Reconfigurable, Physically Composable Hybrid Diffractive Optical Neural Network

Ziang Yin,Yu Yao,Jeff Zhang,Jiaqi Gu
2024-11-09
Abstract:Diffractive optical neural networks (DONNs), leveraging free-space light wave propagation for ultra-parallel, high-efficiency computing, have emerged as promising artificial intelligence (AI) accelerators. However, their inherent lack of reconfigurability due to fixed optical structures post-fabrication hinders practical deployment in the face of dynamic AI workloads and evolving applications. To overcome this challenge, we introduce, for the first time, a multi-dimensional reconfigurable hybrid diffractive ONN system (MDR-HDONN), a physically composable architecture that unlocks a new degree of freedom and unprecedented versatility in DONNs. By leveraging full-system learnability, MDR-HDONN repurposes fixed fabricated optical hardware, achieving exponentially expanded functionality and superior task adaptability through the differentiable learning of system variables. Furthermore, MDR-HDONN adopts a hybrid optical/photonic design, combining the reconfigurability of integrated photonics with the ultra-parallelism of free-space diffractive systems. Extensive evaluations demonstrate that MDR-HDONN has digital-comparable accuracy on various task adaptations with 74x faster speed and 194x lower energy. Compared to prior DONNs, MDR-HDONN shows exponentially larger functional space with 5x faster training speed, paving the way for a new paradigm of versatile, composable, hybrid optical/photonic AI computing. We will open-source our codes.
Optics,Artificial Intelligence,Hardware Architecture
What problem does this paper attempt to address?
The problem this paper attempts to address is: how to overcome the lack of reconfigurability in traditional Diffractive Optical Neural Networks (DONNs) after fabrication, thereby enhancing their ability to adapt to dynamic AI workloads and evolving applications. Specifically, the paper proposes a Multi-Dimensional Reconfigurable Hybrid Diffractive Optical Neural Network system (MDR-HDONN) to address this issue through the following innovations: 1. **Multi-Dimensional Reconfigurability**: Introducing multiple learnable system parameters, including light source wavelength (λ), meta-atom pixel size (s), diffraction distance (z), phase mask orientation (O), phase mask arrangement order (P), polarization combination factor (β), etc. These parameters can be optimized during task adaptation. 2. **Physically Composable Architecture**: Designing a physically composable architecture that allows flexible adjustment of phase mask orientation and arrangement order, thereby achieving exponential functional expansion. 3. **Hybrid Optical/Photonic System**: Combining the advantages of free-space diffractive optics and reconfigurable integrated photonics, leveraging the ultra-parallel processing capability of free-space diffractive optics and the high reconfigurability of integrated photonics to achieve joint neural network acceleration. 4. **Automatic Learning Algorithm**: Proposing an automatic learning algorithm that optimizes the aforementioned system parameters through gradient descent, enabling the system to better adapt to new tasks and perform excellently across multiple tasks. Through extensive experimental evaluation, the paper demonstrates that MDR-HDONN significantly outperforms existing DONNs in terms of task adaptation performance, speed, and energy efficiency, paving the way for a new paradigm of multifunctional, composable, hybrid optical/photonic AI computing.