Condensed Matter > Mesoscale and Nanoscale Physics
[Submitted on 11 Nov 2022 (v1), last revised 14 Jun 2024 (this version, v4)]
Title:Neuromorphic Few-Shot Learning: Generalization in Multilayer Physical Neural Networks
View PDF HTML (experimental)Abstract:Physical neuromorphic computing, exploiting the complex dynamics of physical systems, has seen rapid advancements in implementation sophistication and performance. Physical reservoir computing, a subset of neuromorphic computing where only the output layer is trained, faces limitations due to its reliance on single systems. This constrains output dimensionality and dynamic range, limiting performance to a narrow range of tasks and posing a significant challenge to the field's versatility.
We address this issue by engineering a suite of nanomagnetic arrays as physical reservoirs, demonstrating that tuning microstate space and geometry can significantly broaden dynamics and enhance performance. We interconnect these reservoirs in parallel and series, where each node is a distinct system to create a multilayer neural network architecture. The output of one reservoir is recorded, scaled and virtually fed as input to the next reservoir. All data is experimentally measured with outputs combined offline. This networked approach increases output dimensionality, internal dynamics and computational performance. We demonstrate that a physical neuromorphic system is capable of achieving an overparameterised state. This state facilitates meta-learning on small training sets, yielding strong performance across a wide range of tasks. Our approach's efficacy is further demonstrated through few-shot learning, where the system rapidly adapts to new tasks.
Submission history
From: Kilian Stenning [view email][v1] Fri, 11 Nov 2022 17:45:58 UTC (13,379 KB)
[v2] Mon, 20 Feb 2023 17:36:41 UTC (13,379 KB)
[v3] Tue, 30 May 2023 10:39:32 UTC (30,227 KB)
[v4] Fri, 14 Jun 2024 08:33:03 UTC (45,096 KB)
Current browse context:
cond-mat.mes-hall
Change to browse by:
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
IArxiv Recommender
(What is IArxiv?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.