Emergent self-adaptation in an integrated photonic neural network for backpropagation-free learning
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2023
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916493441105920 |
|---|---|
| author | Lugnan, Alessio Aggarwal, Samarth Brückerhoff-Plückelmann, Frank Wright, C. David Pernice, Wolfram H. P. Bhaskaran, Harish Bienstman, Peter |
| author_facet | Lugnan, Alessio Aggarwal, Samarth Brückerhoff-Plückelmann, Frank Wright, C. David Pernice, Wolfram H. P. Bhaskaran, Harish Bienstman, Peter |
| contents | Plastic self-adaptation, nonlinear recurrent dynamics and multi-scale memory are desired features in hardware implementations of neural networks, because they enable them to learn, adapt and process information similarly to the way biological brains do. In this work, we experimentally demonstrate these properties occurring in arrays of photonic neurons. Importantly, this is realised autonomously in an emergent fashion, without the need for an external controller setting weights and without explicit feedback of a global reward signal. Using a hierarchy of such arrays coupled to a backpropagation-free training algorithm based on simple logistic regression, we are able to achieve a performance of 98.2% on the MNIST task, a popular benchmark task looking at classification of written digits. The plastic nodes consist of silicon photonics microring resonators covered by a patch of phase-change material that implements nonvolatile memory. The system is compact, robust, and straightforward to scale up through the use of multiple wavelengths. Moreover, it constitutes a unique platform to test and efficiently implement biologically plausible learning schemes at a high processing speed. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2312_03802 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Emergent self-adaptation in an integrated photonic neural network for backpropagation-free learning Lugnan, Alessio Aggarwal, Samarth Brückerhoff-Plückelmann, Frank Wright, C. David Pernice, Wolfram H. P. Bhaskaran, Harish Bienstman, Peter Optics Plastic self-adaptation, nonlinear recurrent dynamics and multi-scale memory are desired features in hardware implementations of neural networks, because they enable them to learn, adapt and process information similarly to the way biological brains do. In this work, we experimentally demonstrate these properties occurring in arrays of photonic neurons. Importantly, this is realised autonomously in an emergent fashion, without the need for an external controller setting weights and without explicit feedback of a global reward signal. Using a hierarchy of such arrays coupled to a backpropagation-free training algorithm based on simple logistic regression, we are able to achieve a performance of 98.2% on the MNIST task, a popular benchmark task looking at classification of written digits. The plastic nodes consist of silicon photonics microring resonators covered by a patch of phase-change material that implements nonvolatile memory. The system is compact, robust, and straightforward to scale up through the use of multiple wavelengths. Moreover, it constitutes a unique platform to test and efficiently implement biologically plausible learning schemes at a high processing speed. |
| title | Emergent self-adaptation in an integrated photonic neural network for backpropagation-free learning |
| topic | Optics |
| url | https://arxiv.org/abs/2312.03802 |