Emergent self-adaptation in an integrated photonic neural network for backpropagation-free learning DOI Creative Commons
Alessio Lugnan, Samarth Aggarwal, Frank Brückerhoff‐Plückelmann

et al.

arXiv (Cornell University), Journal Year: 2023, Volume and Issue: unknown

Published: Jan. 1, 2023

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 process information similarly the way biological brains do. In this work, we experimentally demonstrate these properties occurring arrays photonic neurons. Importantly, is realised autonomously an emergent fashion, without need for external controller setting weights explicit feedback a global reward signal. Using hierarchy such coupled backpropagation-free training algorithm based on simple logistic regression, able achieve performance 98.2% MNIST task, popular benchmark task looking at classification written digits. The plastic nodes consist silicon photonics microring resonators covered by patch phase-change material that implements nonvolatile memory. system compact, robust, straightforward scale up through use multiple wavelengths. Moreover, it constitutes unique platform test efficiently implement biologically plausible learning schemes high processing speed.

Language: Английский

Neuromorphic Computing for Smart Agriculture DOI Creative Commons
Siyi Lu, Xinqing Xiao

Agriculture, Journal Year: 2024, Volume and Issue: 14(11), P. 1977 - 1977

Published: Nov. 4, 2024

Neuromorphic computing has received more and attention recently since it can process information interact with the world like human brain. Agriculture is a complex system that includes many processes of planting, breeding, harvesting, processing, storage, logistics, consumption. Smart devices in association artificial intelligence (AI) robots Internet Things (IoT) systems have been used also need to be improved accommodate growth computing. great potential promote development smart agriculture. The aim this paper describe current principles neuromorphic technology, explore examples applications agriculture, consider future route synapses, neurons, neural networks (ANNs). A expected improve agricultural production efficiency ensure food quality safety for nutrition health agriculture future.

Language: Английский

Citations

14
A general framework for interpretable neural learning based on local information-theoretic goal functions DOI Creative Commons
Abdullah Makkeh, Marcel Graetz, Andreas C. Schneider

et al.

Proceedings of the National Academy of Sciences, Journal Year: 2025, Volume and Issue: 122(10)

Published: March 5, 2025

Despite the impressive performance of biological and artificial networks, an intuitive understanding how their local learning dynamics contribute to network-level task solutions remains a challenge this date. Efforts bring more scale indeed lead valuable insights, however, general constructive approach describe goals that is both interpretable adaptable across diverse tasks still missing. We have previously formulated information processing goal highly for model neuron with compartmental structure. Building on recent advances in Partial Information Decomposition (PID), we here derive corresponding parametric rule, which allows us introduce "infomorphic" neural networks. demonstrate versatility these networks perform from supervised, unsupervised, memory learning. By leveraging nature PID framework, infomorphic represent tool advance our intricate structure

Language: Английский

Citations

0
Organoid Intelligence: Bridging Artificial Intelligence for Biological Computing and Neurological Insights DOI Creative Commons
Sangeeta Ballav, Amit Ranjan,

Shubhayan Sur

et al.

Biochemistry, Journal Year: 2024, Volume and Issue: unknown

Published: March 8, 2024

Brain organoid implications have opened vast avenues in the realm of interdisciplinary research, particularly growing field intelligence (OI). A brain is a three-dimensional (3D), lab-grown structure that mimics certain aspects human organization and function. The integration technology with computational methods to enhance understanding behavior predict their responses various stimuli known as OI. ability organoids adapt memorize, key area exploration. OI encapsulates confluence breakthroughs stem cell technology, bioengineering, artificial (AI). This chapter delves deep into myriad potentials OI, encompassing an enhanced cognitive functions, achieving significant biological proficiencies. Such advancements stand offer unique complementarity conventional computing methods. sphere signify transformative stride towards more intricate grasp its multifaceted intricacies. intersection biology machine learning rapidly evolving reshaping our life health. convergence driving numerous areas, including genomics, drug discovery, personalized medicine, synthetic biology.

Language: Английский

Citations

2
Brainwave implanted reservoir computing DOI Creative Commons
Liyu Chen, Yi-Chun Chen, Jason C. Huang

et al.

AIP Advances, Journal Year: 2024, Volume and Issue: 14(1)

Published: Jan. 1, 2024

This work aims to build a reservoir computing system recognize signals with the help of brainwaves as input signals. The brainwave were acquired participants listening human brain in this study can be regarded assistant neural networks or non-linear activation function improve signal recognition. We showed that within frequency ranges from 14 16, 20, 30, and 32 Hz, mean squared errors recognition lower than those without brainwaves. result has demonstrated responses obtain more precise results.

Language: Английский

Citations

1
Evidence of interrelated cognitive-like capabilities in large language models: Indications of artificial general intelligence or achievement? DOI Creative Commons

David Ilić,

Gilles E. Gignac

Intelligence, Journal Year: 2024, Volume and Issue: 106, P. 101858 - 101858

Published: Aug. 29, 2024

Language: Английский

Citations

1
Computing with oscillators from theoretical underpinnings to applications and demonstrators DOI Creative Commons
Aida Todri‐Sanial, Corentin Delacour, Madeleine Abernot

et al.

Deleted Journal, Journal Year: 2024, Volume and Issue: 1(1)

Published: Dec. 4, 2024

Networks of coupled oscillators have far-reaching implications across various fields, providing insights into a plethora dynamics. This review offers an in-depth overview computing with covering computational capability, synchronization occurrence and mathematical formalism. We discuss numerous circuit design implementations, technology choices applications from pattern retrieval, combinatorial optimization problems to machine learning algorithms. also outline perspectives broaden the understanding oscillator

Language: Английский

Citations

1
Emergent self-adaptation in an integrated photonic neural network for backpropagation-free learning DOI Creative Commons
Alessio Lugnan, Samarth Aggarwal, Frank Brückerhoff‐Plückelmann

et al.

Research Square (Research Square), Journal Year: 2023, Volume and Issue: unknown

Published: Nov. 28, 2023

Abstract 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 process information similarly the way biological brains do. In this work, we experimentally demonstrate these properties occurring arrays photonic neurons. Importantly, is realised autonomously an emergent fashion, without need for external controller setting weights explicit feedback a global reward signal. Using hierarchy such coupled backpropagation-free training algorithm based on simple logistic regression, able achieve performance 98.2% MNIST task, popular benchmark task looking at classification written digits. The plastic nodes consist silicon photonics microring resonators covered by patch phase-change material that implements nonvolatile memory. system compact, robust, straightforward scale up through use multiple wavelengths. Moreover, it constitutes unique platform test efficiently implement biologically plausible learning schemes high processing speed.

Language: Английский

Citations

2
Emergent self-adaptation in an integrated photonic neural network for backpropagation-free learning DOI Creative Commons
Alessio Lugnan, Samarth Aggarwal, Frank Brückerhoff‐Plückelmann

et al.

arXiv (Cornell University), Journal Year: 2023, Volume and Issue: unknown

Published: Jan. 1, 2023

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 process information similarly the way biological brains do. In this work, we experimentally demonstrate these properties occurring arrays photonic neurons. Importantly, is realised autonomously an emergent fashion, without need for external controller setting weights explicit feedback a global reward signal. Using hierarchy such coupled backpropagation-free training algorithm based on simple logistic regression, able achieve performance 98.2% MNIST task, popular benchmark task looking at classification written digits. The plastic nodes consist silicon photonics microring resonators covered by patch phase-change material that implements nonvolatile memory. system compact, robust, straightforward scale up through use multiple wavelengths. Moreover, it constitutes unique platform test efficiently implement biologically plausible learning schemes high processing speed.

Language: Английский

Citations

0