Digital-analog hybrid matrix multiplication processor for optical neural networks (2401.15061v1)
Abstract: The computational demands of modern AI have spurred interest in optical neural networks (ONNs) which offer the potential benefits of increased speed and lower power consumption. However, current ONNs face various challenges,most significantly a limited calculation precision (typically around 4 bits) and the requirement for high-resolution signal format converters (digital-to-analogue conversions (DACs) and analogue-to-digital conversions (ADCs)). These challenges are inherent to their analog computing nature and pose significant obstacles in practical implementation. Here, we propose a digital-analog hybrid optical computing architecture for ONNs, which utilizes digital optical inputs in the form of binary words. By introducing the logic levels and decisions based on thresholding, the calculation precision can be significantly enhanced. The DACs for input data can be removed and the resolution of the ADCs can be greatly reduced. This can increase the operating speed at a high calculation precision and facilitate the compatibility with microelectronics. To validate our approach, we have fabricated a proof-of-concept photonic chip and built up a hybrid optical processor (HOP) system for neural network applications. We have demonstrated an unprecedented 16-bit calculation precision for high-definition image processing, with a pixel error rate (PER) as low as $1.8\times10{-3}$ at an signal-to-noise ratio (SNR) of 18.2 dB. We have also implemented a convolutional neural network for handwritten digit recognition that shows the same accuracy as the one achieved by a desktop computer. The concept of the digital-analog hybrid optical computing architecture offers a methodology that could potentially be applied to various ONN implementations and may intrigue new research into efficient and accurate domain-specific optical computing architectures for neural networks.
- Wu, T. et al. A brief overview of chatgpt: The history, status quo and potential future development. IEEE/CAA Journal of Automatica Sinica 10, 1122–1136 (2023).
- Review of ai and compute. https://openai.com/blog/ai-and-compute/ (2018).
- Jouppi, N. P. et al. In-datacenter performance analysis of a tensor processing unit. In Proceedings of the 44th annual international symposium on computer architecture, 1–12 (2017).
- Merolla, P. A. et al. A million spiking-neuron integrated circuit with a scalable communication network and interface. Science 345, 668–673 (2014).
- The end of moore’s law: A new beginning for information technology. Computing in Science Engineering 19, 41–50 (2017).
- Rupp, K. et al. 42 years of microprocessor trend data. https://www.karlrupp.net/ 2018/02/42-years-of-microprocessor-trend-data/ [Online] (2018).
- Science and research policy at the end of moore’s law. Nature Electronics 1, 14–21 (2018).
- Why future supercomputing requires optics. Nature Photonics 4, 261–263 (2010).
- Miller, D. A. B. Perfect optics with imperfect components. Optica 2, 747–750 (2015).
- Neuromorphic photonic integrated circuits. IEEE Journal of Selected Topics in Quantum Electronics 24, 1–15 (2018).
- Shen, Y. et al. Deep learning with coherent nanophotonic circuits. Nature Photonics 11, 441–446 (2017).
- Feldmann, J. et al. Parallel convolutional processing using an integrated photonic tensor core. Nature 589, 52–58 (2021).
- Nahmias, M. A. et al. Photonic multiply-accumulate operations for neural networks. IEEE Journal of Selected Topics in Quantum Electronics 26, 1–18 (2019).
- Multi-channel control for microring weight banks. Optics Express 24, 8895–8906 (2016).
- Tait, A. N. et al. Neuromorphic photonic networks using silicon photonic weight banks. Scientific reports 7, 1–10 (2017).
- Zhang, W. et al. Silicon microring synapses enable photonic deep learning beyond 9-bit precision. Optica 9, 579–584 (2022).
- Bai, B. et al. Microcomb-based integrated photonic processing unit. Nature Communications 14, Article number: 66 (2023).
- Deep learning with limited numerical precision. In International conference on machine learning, 1737–1746 (PMLR, 2015).
- Murmann, B. ADC Performance Survey 1997-2023. [Online]. Available: https://github.com/bmurmann/ADC-survey.
- Van der Walt, S. et al. scikit-image: image processing in python. PeerJ 2, e453 (2014).
- Kong, D. et al. Intra-datacenter interconnects with a serialized silicon optical frequency comb modulator. Journal of Lightwave Technology 38, 4677–4682 (2020).
- Ten-channel discrete multi-tone modulation using silicon microring modulator array. In 2016 Optical Fiber Communications Conference and Exhibition (OFC), 1–3 (2016).
- Kong, D. et al. 100 Gbit/s PAM-16 Transmission in the 2-μ𝜇\muitalic_μm Band over a 1.15-km Hollow-Core Fiber. In Optical Fiber Communications Conference (OFC) 2021, Th4E.6 (IEEE, 2021).
- The MNIST database of handwritten digits. [Online]. Available: http://yann.lecun.com/exdb/mnist/.
- 640 gbits/s photonic logic gates. Opt. Lett. 35, 3955–3957 (2010).
- Xingyuan, X. et al. 11 TOPS photonic convolutional accelerator for optical neural networks. Nature 589, 44–51 (2021).
- Luo, X. et al. Metasurface-enabled on-chip multiplexed diffractive neural networks in the visible. Light: Science & Applications 11, Article number: 158 (2022).
- Luo, Y. et al. Design of task-specific optical systems using broadband diffractive neural networks. Light: Science & Applications 8, Article number: 112 (2019).
- Electro-optic frequency combs. Adv. Opt. Photon. 12, 223–287 (2020).
- Kong, D. et al. Cavity-less sub-picosecond pulse generation for the demultiplexing of a 640 Gbaud OTDM signal. In IEEE Photonics Conference (IPC) 2015, WG1.2 (IEEE, 2015).
- Dong, P. et al. Wavelength-tunable silicon microring modulator. Opt. Express 18, 10941–10946 (2010).