Filters








104,207 Hits in 4.1 sec

Making Neural Networks FAIR [article]

Anna Nguyen and Tobias Weller and Michael Färber and York Sure-Vetter
2020 arXiv   pre-print
In this paper, we first present the neural network ontology FAIRnets Ontology, an ontology to make existing neural network models findable, accessible, interoperable, and reusable according to the FAIR  ...  neural networks to data scientists.  ...  Our approach aims to consider such neural networks and make them available as FAIR data. CONCLUSION is paper was dedicated to making neural networks FAIR.  ... 
arXiv:1907.11569v4 fatcat:3pbgjqst6bg2jkcpwfq6x6anjm

COVID-Net Clinical ICU: Enhanced Prediction of ICU Admission for COVID-19 Patients via Explainability and Trust Quantification [article]

Audrey Chung, Mahmoud Famouri, Andrew Hryniowski, Alexander Wong
2021 arXiv   pre-print
Motivated by the need for transparent and trustworthy ICU admission clinical decision support, we introduce COVID-Net Clinical ICU, a neural network for ICU admission prediction based on patient clinical  ...  By digging deeper into when and why clinical predictive models makes certain decisions, we can uncover key factors in decision making for critical clinical decision support tasks such as ICU admission  ...  The neural network generally behaves in a fair manner for both gender and age demographics.  ... 
arXiv:2109.06711v2 fatcat:rtgzm2dv3bghrkqw2obehwqgsa

Fair Interpretable Learning via Correction Vectors [article]

Mattia Cerrato and Marius Köppel and Alexander Segner and Stefan Kramer
2022 arXiv   pre-print
Neural network architectures have been extensively employed in the fair representation learning setting, where the objective is to learn a new representation for a given vector which is independent of  ...  However, as neural networks are inherently opaque, these methods are hard to comprehend, which limits their usefulness.  ...  One issue with fair representation learning algorithms based on neural networks is their opaqueness.  ... 
arXiv:2201.06343v1 fatcat:s7kbxinwsva2zfjoqy5eymwnvq

Causality-based Neural Network Repair [article]

Bing Sun, Jun Sun, Hong Long Pham, Jie Shi
2022 arXiv   pre-print
Similar to traditional decision-making programs, neural networks can have defects that need to be repaired. The defects may cause unsafe behaviors, raise security concerns or unjust societal impacts.  ...  Our experiment results show that CARE is able to repair all neural networks efficiently and effectively. For fairness repair tasks, CARE successfully improves fairness by 61.91% on average.  ...  Similar to traditional decision-making programs, neural networks inevitably have defects and need to be repaired at times.  ... 
arXiv:2204.09274v2 fatcat:4izca4dgr5bjffcdsu54nirh2e

PREDICTING STUDENTS«¤?? GRADE SCORES USING TRAINING FUNCTIONS OF ARTIFICIAL NEURAL NETWORK

B. A. KALEJAYE, O. FOLORUNSO, O. L. USMAN
2016 Journal of Natural Sciences, Engineering and Technology  
CGPA values simulated by the network are compared with the actual final CGPA to determine the efficacy of each of the three feed-forward neural networks used.  ...  This study focuses on the use of artificial neural network (ANN) model for predicting students«¤?? academic performance in a University System, based on the previous datasets.  ...  Wang and Mitrovic (2002) applied neural networks to forecast the number of errors that a trainee will make using problem-specific attributes and the trainees' current level as input variables.  ... 
doi:10.51406/jnset.v14i1.1482 fatcat:7mirdxkggnae7jhxhbapow35om

Gradient Reversal Against Discrimination [article]

Edward Raff, Jared Sylvester
2018 arXiv   pre-print
No methods currently exist for making arbitrary neural networks fair.  ...  In this work we introduce GRAD, a new and simplified method to producing fair neural networks that can be used for auto-encoding fair representations or directly with predictive networks.  ...  This makes it the only neural network-based approach to fairness that offers both task flexibility and specificity.  ... 
arXiv:1807.00392v1 fatcat:ms5btkpqx5defohngqwvkx53nu

Leveraging Semi-Supervised Learning for Fairness using Neural Networks [article]

Vahid Noroozi, Sara Bahaadini, Samira Sheikhi, Nooshin Mojab, Philip S. Yu
2019 arXiv   pre-print
In this paper, we propose a semi-supervised algorithm using neural networks benefiting from unlabeled data to not just improve the performance but also improve the fairness of the decision-making process  ...  There has been a growing concern about the fairness of decision-making systems based on machine learning.  ...  [21] which is a model based on neural networks to address the fairness problem.  ... 
arXiv:1912.13230v1 fatcat:rxlkti454zcotbzclerbc27hsa

Intra-Processing Methods for Debiasing Neural Networks [article]

Yash Savani, Colin White, Naveen Sundar Govindarajulu
2020 arXiv   pre-print
All of our techniques can be used for all popular group fairness measures such as equalized odds or statistical parity difference.  ...  The need for neural networks.  ...  Next, we present three new optimization-based techniques for post-hoc debiasing of neural networks, each of which work for any group fairness measure.  ... 
arXiv:2006.08564v2 fatcat:pauvm5izljcmldstpfhlbb3gda

A Comparison of Reinforcement Learning Algorithms in Fairness-Oriented OFDMA Schedulers

Comsa, Zhang, Aydin, Kuonen, Trestian, Ghinea
2019 Information  
This framework makes use of feed forward neural networks to map momentarystates to proper parameterization decisions for the proportional fair scheduler.  ...  Due to large-scale control problems in 5G access networks, the complexity of radioresource management is expected to increase significantly.  ...  This value is reinforced and the controller updates the neural networks in order to improve its decision-making.  ... 
doi:10.3390/info10100315 fatcat:pho2jfu7crhvnbaeca5ly2mkhi

Editorial: The Neural Basis of Human Prosocial Behavior

Yefeng Chen, Hang Ye, Chao Liu, Qi Li
2019 Frontiers in Psychology  
Zheng et al. summarize models of the emotional influence on fairness-related decision making and the corresponding behavioral and neural evidence.  ...  They demonstrated that the fairness-related decision-making processes are context-dependent and are modulated by social support.  ... 
doi:10.3389/fpsyg.2019.02058 pmid:31551889 pmcid:PMC6747013 fatcat:54kx2l43hrf5rh4uxn7tcmep2e

Insights into Fairness through Trust: Multi-scale Trust Quantification for Financial Deep Learning [article]

Alexander Wong, Andrew Hryniowski, Xiao Yu Wang
2020 arXiv   pre-print
More specifically, we conduct multi-scale trust quantification on a deep neural network for the purpose of credit card default prediction to study: 1) the overall trustworthiness of the model 2) the trust  ...  the fairness of models.  ...  By studying trustworthiness of a deep neural network across multiple scales, one can gain deeper insights into not just how trustworthy a deep neural network is, but also where trust breaks down.  ... 
arXiv:2011.01961v1 fatcat:do3dyipu4jah5lfvj6ek4cxule

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

Xuanqi Gao, Juan Zhai, Shiqing Ma, Chao Shen, Yufei Chen, Qian Wang
2022 arXiv   pre-print
With Deep Neural Network (DNN) being integrated into a growing number of critical systems with far-reaching impacts on society, there are increasing concerns on their ethical performance, such as fairness  ...  Comparing with state-of-the-art methods, our approach is lightweight, making it scalable and more efficient.  ...  Neural Network Slicing In Neural Network Slicing, we try to find paths and neurons that contain the optimizer finds contradictory optimization directions for accuracy and fairness.  ... 
arXiv:2204.02567v2 fatcat:e5xaax2jbbecrbzhqg7aoqcrx4

Student Academic Performance Prediction using Artificial Neural Networks: A Case Study

Mubarak Albarka
2019 International Journal of Computer Applications  
This study presents a neural network model capable of predicting student's GPA using students' personal information, academic information, and place of residence.  ...  A sample of 61 Computer Networking students' dataset was used to train and test the model in WEKA software tool. The accuracy of the model was measured using well-known evaluation criteria.  ...  The good results of applying neural networks in prediction and classification problems makes it appropriate for this study.  ... 
doi:10.5120/ijca2019919387 fatcat:axa62gsulngohjpaisodukzunm

Probabilistic Verification of Neural Networks Against Group Fairness [article]

Bing Sun, Jun Sun, Ting Dai, Lijun Zhang
2021 arXiv   pre-print
Fairness is crucial for neural networks which are used in applications with important societal implication.  ...  In this work, we propose an approach to formally verify neural networks against fairness, with a focus on independence-based fairness such as group fairness.  ...  Fairness issues in neural networks are often more 'hidden' than those of traditional decision-making software programs since it is still an open problem on how to interpret neural networks.  ... 
arXiv:2107.08362v1 fatcat:mactwovasncl7nhdxq6jgk6t54

FETA: Fairness Enforced Verifying, Training, and Predicting Algorithms for Neural Networks [article]

Kiarash Mohammadi, Aishwarya Sivaraman, Golnoosh Farnadi
2022 arXiv   pre-print
Algorithmic decision making driven by neural networks has become very prominent in applications that directly affect people's quality of life.  ...  However, such a translation does not always guarantee fair predictions of the trained neural network model.  ...  Hence, we will assume f θ is a ReLU neural network. Formal properties of neural networks are often verified by encoding the semantics of neural networks (f θ ) as logical constraints.  ... 
arXiv:2206.00553v1 fatcat:d6aj65n35vgh5cam362qxinzp4
« Previous Showing results 1 — 15 out of 104,207 results