Bias Neuron

Bias neurons are specific neural network units identified as contributing disproportionately to unfair or biased model outputs across various applications, including speech recognition, natural language processing, and image classification. Current research focuses on detecting these neurons using techniques like integrated gradient methods and mitigating their effects through methods such as neuron suppression, bias potential manipulation, and data preprocessing (e.g., sketching). Understanding and addressing bias neurons is crucial for building fairer and more trustworthy AI systems, impacting fields ranging from healthcare to social justice by reducing algorithmic discrimination and improving model reliability.

Papers

July 14, 2024

June 14, 2024

The Devil is in the Neurons: Interpreting and Mitigating Social Biases in Pre-trained Language Models
Yan Liu, Yu Liu, Xiaokang Chen, Pin-Yu Chen, Daoguang Zan, Min-Yen Kan, Tsung-Yi Ho
Pre Trained Language Model Social Bias Single Neuron Level Trading Devil Machine Translated Bias Neuron

April 22, 2024

Unveiling and Mitigating Generalized Biases of DNNs through the Intrinsic Dimensions of Perceptual Manifolds
Yanbiao Ma, Licheng Jiao, Fang Liu, Lingling Li, Wenping Ma, Shuyuan Yang, Xu Liu, Puhua Chen
Neural Network Deep Neural Network Intrinsic Dimension Bias Neuron

February 8, 2024

Hidden in Plain Sight: Undetectable Adversarial Bias Attacks on Vulnerable Patient Populations
Pranav Kulkarni, Andrew Chan, Nithya Navarathna, Skylar Chan, Paul H. Yi, Vishwa S. Parekh
Hidden Knowledge Plain Sight Prediction Bias Vulnerable Population Bias Neuron Adversarial Bias

January 13, 2024

Accelerated Sampling of Rare Events using a Neural Network Bias Potential
Xinru Hua, Rasool Ahmad, Jose Blanchet, Wei Cai
Importance Sampling Molecular Dynamic Rare Event Monte Carlo Sampling Computational Physic Rare Transition Bias Neuron Accelerated Sampling

November 16, 2023

Mitigating Biases for Instruction-following Language Models via Bias Neurons Elimination
Nakyeong Yang, Taegwan Kang, Jungkyu Choi, Honglak Lee, Kyomin Jung
Language Model Zero Shot Bias Mitigation Mitigating Bias Bias Neuron

April 27, 2023

FLAC: Fairness-Aware Representation Learning by Suppressing Attribute-Class Associations
Ioannis Sarridis, Christos Koutlis, Symeon Papadopoulos, Christos Diou
Representation Learning Fair Representation Biased Datasets Label Bias Biased Classifier Bias Neuron Class Relevant

November 21, 2022

Diffusion Denoising Process for Perceptron Bias in Out-of-distribution Detection
Luping Liu, Yi Ren, Xize Cheng, Rongjie Huang, Chongxuan Li, Zhou Zhao
Distribution Detection Denoising Diffusion Discriminative Model Bi Discriminator Bias Neuron

October 31, 2022

Improving Fairness in Image Classification via Sketching
Ruichen Yao, Ziteng Cui, Xiaoxiao Li, Lin Gu
Deep Neural Network Image Classification Model Fairness Fairness Improvement Sketch to Image Source Bias Bias Neuron

December 25, 2021

NeuronFair: Interpretable White-Box Fairness Testing through Biased Neuron Identification
Haibin Zheng, Zhiqing Chen, Tianyu Du, Xuhong Zhang, Yao Cheng, Shouling Ji, Jingyi Wang, Yue Yu, Jinyin Chen
Neural Network Deep Neural Network FAIrness Testing Bias Neuron

Bias Neuron

Papers

Evolved Developmental Artificial Neural Networks for Multitasking with Advanced Activity Dependence

Improving Neural Biasing for Contextual Speech Recognition by Early Context Injection and Text Perturbation

The Devil is in the Neurons: Interpreting and Mitigating Social Biases in Pre-trained Language Models

Unveiling and Mitigating Generalized Biases of DNNs through the Intrinsic Dimensions of Perceptual Manifolds

Hidden in Plain Sight: Undetectable Adversarial Bias Attacks on Vulnerable Patient Populations

Accelerated Sampling of Rare Events using a Neural Network Bias Potential

Mitigating Biases for Instruction-following Language Models via Bias Neurons Elimination

FLAC: Fairness-Aware Representation Learning by Suppressing Attribute-Class Associations

Diffusion Denoising Process for Perceptron Bias in Out-of-distribution Detection

Improving Fairness in Image Classification via Sketching

NeuronFair: Interpretable White-Box Fairness Testing through Biased Neuron Identification