# L2ight: Enabling On-Chip Learning for Optical Neural Networks via Efficient in-situ Subspace Optimization

arXiv: https://arxiv.org/abs/2110.14807 · HTML: https://mlsanity.com/paper/2110.14807/

## Key facts

- Cited by 21 papers in the map (data through 14 Oct 2026).
- Builds on 30 of its 62 references that are in the map.
- First posted on arXiv 27 Oct 2021.

## Builds on (most cited first, 30 of 30)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms](https://mlsanity.com/paper/1708.07747/) (2017)
- [Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding](https://mlsanity.com/paper/1510.00149/) (2015)
- [Stochastic First- and Zeroth-order Methods for Nonconvex Stochastic Programming](https://mlsanity.com/paper/1309.5549/) (2013)
- [signSGD: Compressed Optimisation for Non-Convex Problems](https://mlsanity.com/paper/1802.04434/) (2018)
- [Learning Structured Sparsity in Deep Neural Networks](https://mlsanity.com/paper/1608.03665/) (2016)
- [Universal Sentence Encoder](https://mlsanity.com/paper/1803.11175/) (2018)
- [Exploiting Linear Structure Within Convolutional Networks for Efficient Evaluation](https://mlsanity.com/paper/1404.0736/) (2014)
- [Direct Feedback Alignment Provides Learning in Deep Neural Networks](https://mlsanity.com/paper/1609.01596/) (2016)
- [Speeding-up Convolutional Neural Networks Using Fine-tuned CP-Decomposition](https://mlsanity.com/paper/1412.6553/) (2014)
- [Scalable Methods for 8-bit Training of Neural Networks](https://mlsanity.com/paper/1805.11046/) (2018)
- [Photonics for artificial intelligence and neuromorphic computing](https://mlsanity.com/paper/2011.00111/) (2020)
- [Convolutional neural networks with low-rank regularization](https://mlsanity.com/paper/1511.06067/) (2015)
- [Highly Scalable Deep Learning Training System with Mixed-Precision: Training ImageNet in Four Minutes](https://mlsanity.com/paper/1807.11205/) (2018)
- [CirCNN: Accelerating and Compressing Deep Neural Networks Using Block-CirculantWeight Matrices](https://mlsanity.com/paper/1708.08917/) (2017)
- [Training of photonic neural networks through in situ backpropagation](https://mlsanity.com/paper/1805.09943/) (2018)
- [Sparse Weight Activation Training](https://mlsanity.com/paper/2001.01969/) (2020)
- [A Comprehensive Linear Speedup Analysis for Asynchronous Stochastic Parallel Optimization from Zeroth-Order to First-Order](https://mlsanity.com/paper/1606.00498/) (2016)
- [meProp: Sparsified Back Propagation for Accelerated Deep Learning with Reduced Overfitting](https://mlsanity.com/paper/1706.06197/) (2017)
- [Stochastic Three Points Method for Unconstrained Smooth Minimization](https://mlsanity.com/paper/1902.03591/) (2019)
- [Efficient On-Chip Learning for Optical Neural Networks Through Power-Aware Sparse Zeroth-Order Optimization](https://mlsanity.com/paper/2012.11148/) (2020)
- [Randomized Automatic Differentiation](https://mlsanity.com/paper/2007.10412/) (2020)
- [Massively Parallel Amplitude-Only Fourier Neural Network](https://mlsanity.com/paper/2008.05853/) (2020)
- [Faster Neural Network Training with Approximate Tensor Operations](https://mlsanity.com/paper/1805.08079/) (2018)
- [Dithered backprop: A sparse and quantized backpropagation algorithm for more efficient deep neural network training](https://mlsanity.com/paper/2004.04729/) (2020)
- [A Stochastic Derivative-Free Optimization Method with Importance Sampling: Theory and Learning to Control](https://mlsanity.com/paper/1902.01272/) (2019)
- [Efficient training and design of photonic neural network through neuroevolution](https://mlsanity.com/paper/1908.08012/) (2019)
- [Efficient Recurrent Neural Networks using Structured Matrices in FPGAs](https://mlsanity.com/paper/1803.07661/) (2018)
- [Subspace Network: Deep Multi-Task Censored Regression for Modeling Neurodegenerative Diseases](https://mlsanity.com/paper/1802.06516/) (2018)

## Built on by (most cited first, 21 of 21)

- [DeepZero: Scaling up Zeroth-Order Optimization for Deep Model Training](https://mlsanity.com/paper/2310.02025/) (2023)
- [Training of Physical Neural Networks](https://mlsanity.com/paper/2406.03372/) (2024)
- [NeurOLight: A Physics-Agnostic Neural Operator Enabling Parametric Photonic Device Simulation](https://mlsanity.com/paper/2209.10098/) (2022)
- [Tensor-Compressed Back-Propagation-Free Training for (Physics-Informed) Neural Networks](https://mlsanity.com/paper/2308.09858/) (2023)
- [Optical Transformers](https://mlsanity.com/paper/2302.10360/) (2023)
- [Dual adaptive training of photonic neural networks](https://mlsanity.com/paper/2212.06141/) (2022)
- [ZeroFlow: Overcoming Catastrophic Forgetting is Easier than You Think](https://mlsanity.com/paper/2501.01045/) (2025)
- [SimPhony: A Device-Circuit-Architecture Cross-Layer Modeling and Simulation Framework for Heterogeneous Electronic-Photonic AI System](https://mlsanity.com/paper/2411.13715/) (2024)
- [SCATTER: Algorithm-Circuit Co-Sparse Photonic Accelerator with Thermal-Tolerant, Power-Efficient In-situ Light Redistribution](https://mlsanity.com/paper/2407.05510/) (2024)
- [Poor Man's Training on MCUs: A Memory-Efficient Quantized Back-Propagation-Free Approach](https://mlsanity.com/paper/2411.05873/) (2024)
- [Integrated multi-operand optical neurons for scalable and hardware-efficient deep learning](https://mlsanity.com/paper/2305.19592/) (2023)
- [Real-Time FJ/MAC PDE Solvers via Tensorized, Back-Propagation-Free Optical PINN Training](https://mlsanity.com/paper/2401.00413/) (2023)
- [FLOPS: Forward Learning with OPtimal Sampling](https://mlsanity.com/paper/2410.05966/) (2024)
- [Scalable Back-Propagation-Free Training of Optical Physics-Informed Neural Networks](https://mlsanity.com/paper/2502.12384/) (2025)
- [Model-free Optical Processors using In Situ Reinforcement Learning with Proximal Policy Optimization](https://mlsanity.com/paper/2507.05583/) (2025)
- [DOCTOR: Dynamic On-Chip Temporal Variation Remediation Toward Self-Corrected Photonic Tensor Accelerators](https://mlsanity.com/paper/2403.02688/) (2024)
- [Toward Lifelong-Sustainable Electronic-Photonic AI Systems via Extreme Efficiency, Reconfigurability, and Robustness](https://mlsanity.com/paper/2509.07396/) (2025)
- [Low-rank surrogate modeling and stochastic zero-order optimization for training of neural networks with black-box layers](https://mlsanity.com/paper/2509.15113/) (2025)
- [Efficient training for large-scale optical neural network using an evolutionary strategy and attention pruning](https://mlsanity.com/paper/2505.12906/) (2025)
- [LLM-Guided Neural Architecture Search for Robust Co-Design of Physical Neural Networks](https://mlsanity.com/paper/2606.10294/) (2026)
- [LCAP: Population-Informed Latent Chip Adaptation from Few Output Probes for Photonic Neural Networks](https://mlsanity.com/paper/2609.16823/) (2026)

## Abstract (arXiv)

Silicon-photonics-based optical neural network (ONN) is a promising hardware platform that could represent a paradigm shift in efficient AI with its CMOS-compatibility, flexibility, ultra-low execution latency, and high energy efficiency. In-situ training on the online programmable photonic chips is appealing but still encounters challenging issues in on-chip implementability, scalability, and efficiency. In this work, we propose a closed-loop ONN on-chip learning framework L2ight to enable scalable ONN mapping and efficient in-situ learning. L2ight adopts a three-stage learning flow that first calibrates the complicated photonic circuit states under challenging physical constraints, then performs photonic core mapping via combined analytical solving and zeroth-order optimization. A subspace learning procedure with multi-level sparsity is integrated into L2ight to enable in-situ gradient evaluation and fast adaptation, unleashing the power of optics for real on-chip intelligence. Extensive experiments demonstrate our proposed L2ight outperforms prior ONN training protocols with 3-order-of-magnitude higher scalability and over 30X better efficiency, when benchmarked on various models and learning tasks. This synergistic framework is the first scalable on-chip learning solution that pushes this emerging field from intractable to scalable and further to efficient for next-generation self-learnable photonic neural chips. From a co-design perspective, L2ight also provides essential insights for hardware-restricted unitary subspace optimization and efficient sparse training. We open-source our framework at https://github.com/JeremieMelo/L2ight.