Analog Neural Network

Analog neural networks (ANNs) aim to leverage the energy efficiency and speed of analog circuits for implementing artificial neural networks, addressing the power and scalability limitations of digital approaches. Current research focuses on mitigating the effects of hardware noise through novel regularization techniques and denoising blocks, developing hybrid digital-analog training methods like Equilibrium Propagation, and exploring diverse architectures such as Deep Hopfield Networks and KirchhoffNets. This field is significant for its potential to enable energy-efficient AI at the edge, particularly in applications like medical imaging and robotics, where low-power consumption and fast inference are crucial.

Papers

February 3, 2023

Symbiosis of an artificial neural network and models of biological neurons: training and testing
Tatyana Bogatenko, Konstantin Sergeev, Andrei Slepnev, Jürgen Kurths, Nadezhda Semenova
Neural Network Training Data Full Model Analog Neural Network Artificial Neuron Biological Neuron Symbiotic Transformer

October 14, 2022

AnalogVNN: A fully modular framework for modeling and optimizing photonic neural networks
Vivswan Shah, Nathan Youngblood
Modular Framework Analog Neural Network Photonic Neural Network Neuromorphic Photonic

September 30, 2022

MLPInit: Embarrassingly Simple GNN Training Acceleration with MLP Initialization
Xiaotian Han, Tong Zhao, Yozen Liu, Xia Hu, Neil Shah
Graph Neural Network Message Passing Graph Neural Network GNN Training Analog Neural Network Peer Learning

September 19, 2022

An accurate and flexible analog emulation of AdEx neuron dynamics in silicon
Sebastian Billaudelle, Johannes Weis, Philipp Dauer, Johannes Schemmel
Neuromorphic Hardware Analog Neural Network Synaptic Connectivity BrainScaleS 2 Silicon Atom

August 2, 2022

Analog Gated Recurrent Neural Network for Detecting Chewing Events
Kofi Odame, Maria Nyamukuru, Mohsen Shahghasemi, Shengjie Bi, David Kotz
Neural Network Recurrent Neural Network Analog Neural Network Complementary Metal Oxide Semiconductor Virtual Analog

July 8, 2022

RF-Photonic Deep Learning Processor with Shannon-Limited Data Movement
Ronald Davis, Zaijun Chen, Ryan Hamerly, Dirk Englund
Neural Network DNN Accelerator Analog Neural Network

May 18, 2022

Single-Shot Optical Neural Network
Liane Bernstein, Alexander Sludds, Christopher Panuski, Sivan Trajtenberg-Mills, Ryan Hamerly, Dirk Englund
Neural Network Optical Communication Analog Neural Network Optical Processor Free Space Optical

May 12, 2022

Adaptive Block Floating-Point for Analog Deep Learning Hardware
Ayon Basumallik, Darius Bunandar, Nicholas Dronen, Nicholas Harris, Ludmila Levkova, Calvin McCarter, Lakshmi Nair, David Walter, David Widemann
Quantization Aware Training Analog Neural Network Floating Point Energy Efficient Neural Network Mixed Signal Average Precision Loss

April 21, 2022

MRAM-based Analog Sigmoid Function for In-memory Computing
Md Hasibul Amin, Mohammed Elbtity, Mohammadreza Mohammadi, Ramtin Zand
Memory Computing Analog Neural Network Crossbar Array Memristive Crossbar

April 1, 2022

Physical Deep Learning with Biologically Plausible Training Method
Mitsumasa Nakajima, Katsuma Inoue, Kenji Tanaka, Yasuo Kuniyoshi, Toshikazu Hashimoto, Kohei Nakajima
Neuromorphic Computing Analog Neural Network Physical Neural Network Optoelectronic Loop Neuron

January 5, 2022

Improving Spiking Neural Network Accuracy Using Time-based Neurons
Hanseok Kim, Woo-Seok Choi
Spiking Neural Network Neuromorphic Computing Neuron Model Analog Neural Network Time Dependent Neural

November 18, 2021

Continuous learning of spiking networks trained with local rules
Dmitry Antonov, Kirill Sviatov, Sergey Sukhov
Catastrophic Forgetting Spiking Neural Network Network Programming Continuous Learning Biological Neural Network Spike Timing Dependent Plasticity Analog Neural Network Local Rule

November 10, 2021

AnalogNets: ML-HW Co-Design of Noise-robust TinyML Models and Always-On Analog Compute-in-Memory Accelerator
Chuteng Zhou, Fernando Garcia Redondo, Julian Büchel, Irem Boybat, Xavier Timoneda Comas, S. R. Nandakumar, Shidhartha Das, Abu Sebastian, Manuel Le Gallo, Paul N. Whatmough
TinyML Model Memory Computing Analog Neural Network Phantom 2D Accelerator Analog Compute in Memory Phase Change Memory