Selective Neuron

Selective neurons, specialized nerve cells exhibiting heightened responsiveness to specific stimuli, are a key focus in understanding neural network function, both biological and artificial. Current research investigates how these neurons develop during training, their role in network performance (including their surprising dispensability in some cases), and their potential use in improving fairness and efficiency in artificial neural networks. This research utilizes various model architectures, including convolutional networks and spiking neural networks, employing techniques like activation maximization and ablation studies to probe neuron function and importance. Understanding selective neurons offers crucial insights into brain function and promises to improve the design and interpretability of artificial intelligence systems.

Papers

December 28, 2023

Fast gradient-free activation maximization for neurons in spiking neural networks
Nikita Pospelov, Andrei Chertkov, Maxim Beketov, Ivan Oseledets, Konstantin Anokhin
Neural Network Single Neuron Level Biological Neural Network Artificial Neuron Activation Maximization Discrete Activation Selective Neuron

May 27, 2023

On the special role of class-selective neurons in early training
Omkar Ranadive, Nikhil Thakurdesai, Ari S Morcos, Matthew Leavitt, Stéphane Deny
Supervised ImageNet Integral Role Deep Network Training Dynamic Individual Neuron Selective Engagement Selective Neuron

April 5, 2023

Multi-Domain Norm-referenced Encoding Enables Data Efficient Transfer Learning of Facial Expression Recognition
Michael Stettler, Alexander Lappe, Nick Taubert, Martin Giese
Transfer Learning Facial Expression Facial Expression Recognition Expression Recognition Selective Neuron

November 8, 2022

Much Easier Said Than Done: Falsifying the Causal Relevance of Linear Decoding Methods
Lucas Hayne, Abhijit Suresh, Hunar Jain, Rahul Kumar, R. McKell Carter
Linear Classifier Selective Engagement Decoding Method Causal Influence Selective Neuron Additive Decoder

August 13, 2022

Modeling biological face recognition with deep convolutional neural networks
Leonard E. van Dyck, Walter R. Gruber
Deep Convolutional Neural Network Face Recognition Object Recognition Selective Neuron

April 6, 2022

FairNeuron: Improving Deep Neural Network Fairness with Adversary Games on Selective Neurons
Xuanqi Gao, Juan Zhai, Shiqing Ma, Chao Shen, Yufei Chen, Qian Wang
Neural Network Model Fairness Fairness Accuracy Adversarial Game Fairness Aware Training Selective Neuron

Selective Neuron

Papers

Fast gradient-free activation maximization for neurons in spiking neural networks

On the special role of class-selective neurons in early training

Multi-Domain Norm-referenced Encoding Enables Data Efficient Transfer Learning of Facial Expression Recognition

Much Easier Said Than Done: Falsifying the Causal Relevance of Linear Decoding Methods

Modeling biological face recognition with deep convolutional neural networks

FairNeuron: Improving Deep Neural Network Fairness with Adversary Games on Selective Neurons