Filters








519 Hits in 2.9 sec

A Cognitive Approach to Assessing the Materials in Problem-Based Learning Environments

John R. Drake, Ravi Paul
2021 Journal of Information Technology Education Innovations in Practice  
Aim/Purpose: The purpose of this paper is to develop and evaluate a debiasing-based approach to assessing the learning materials in problem-based learning (PBL) environments.  ...  Recommendation for Researchers: This study uses debiasing theory to improve course techniques.  ...  The debiasing strategies were worded to make sense in the context of the project. For example, instead of "learn about the task domain," the survey said, "learn about e-commerce domain".  ... 
doi:10.28945/4812 fatcat:llrojemrcbe6fpl4nb4iqbinu4

Debiased Recommendation with User Feature Balancing [article]

Mengyue Yang, Guohao Cai, Furui Liu, Zhenhua Dong, Xiuqiang He, Jianye Hao, Jun Wang, Xu Chen
2022 arXiv   pre-print
To the best of our knowledge, this paper is the first work on debiased recommendation based on confounder balancing.  ...  To alleviate these problems, in this paper, we propose a novel debiased recommendation framework based on user feature balancing.  ...  domain.  ... 
arXiv:2201.06056v1 fatcat:2474gyilsfb75iavg4abdiopjm

Skin Deep Unlearning: Artefact and Instrument Debiasing in the Context of Melanoma Classification [article]

Peter J. Bevan, Amir Atapour-Abarghouei
2021 arXiv   pre-print
Contributions of this work include the application of different debiasing techniques for artefact bias removal and the concept of instrument bias unlearning for domain generalisation in melanoma detection  ...  Our experimental results provide evidence that the effects of each of the aforementioned biases are notably reduced, with different debiasing techniques excelling at different tasks.  ...  as the base architecture.  ... 
arXiv:2109.09818v3 fatcat:pk56pg7csjdrrm7unyiq5wkuy4

Learning from Failure: Training Debiased Classifier from Biased Classifier [article]

Junhyun Nam, Hyuntak Cha, Sungsoo Ahn, Jaeho Lee, Jinwoo Shin
2020 arXiv   pre-print
Based on the observations, we propose a failure-based debiasing scheme by training a pair of neural networks simultaneously.  ...  We first observe that neural networks learn to rely on the spurious correlation only when it is "easier" to learn than the desired knowledge, and such reliance is most prominent during the early phase  ...  In this paper, we propose a failure-based debiasing scheme, coined Learning from Failure (LfF).  ... 
arXiv:2007.02561v2 fatcat:d5i6mgy2pzatrixenen4avx54u

Debiasing classifiers: is reality at variance with expectation? [article]

Ashrya Agrawal and Florian Pfisterer and Bernd Bischl and Francois Buet-Golfouse and Srijan Sood and Jiahao Chen and Sameena Shah and Sebastian Vollmer
2021 arXiv   pre-print
A rigorous evaluation of the debiasing treatment effect requires extensive cross-validation beyond what is usually done.  ...  Follow-up experiments validate the theoretical prediction that the estimation variance depends strongly on the base rates of the protected class.  ...  This document is not intended as investment research or investment advice, or a recommendation, offer or solicitation for the purchase or sale of any security, financial instrument, financial product or  ... 
arXiv:2011.02407v2 fatcat:3gckzpqfsbdybp662rmnteth3q

Contrastive Learning for Debiased Candidate Generation in Large-Scale Recommender Systems [article]

Chang Zhou, Jianxin Ma, Jianwei Zhang, Jingren Zhou, Hongxia Yang
2021 arXiv   pre-print
Based on the theoretical discovery, we design CLRec, a contrastive learning method to improve DCG in terms of fairness, effectiveness and efficiency in recommender systems with extremely large candidate  ...  Deep candidate generation (DCG) that narrows down the collection of relevant items from billions to hundreds via representation learning has become prevalent in industrial recommender systems.  ...  Contrastive Learning Contrastive learning, which aims to learn high-quality representations via self-supervised pretext tasks, recently achieves remarkable successes in various domains, e.g., speech processing  ... 
arXiv:2005.12964v9 fatcat:ju2bph7nqbfhhlvyswuexddova

Information Theoretic Counterfactual Learning from Missing-Not-At-Random Feedback [article]

Zifeng Wang and Xi Chen and Rui Wen and Shao-Lun Huang and Ercan E. Kuruoglu and Yefeng Zheng
2020 arXiv   pre-print
between the factual and counterfactual domains.  ...  Missing-at-random (MAR) data, namely randomized controlled trials (RCTs), are usually required by most previous counterfactual learning methods for debiasing learning.  ...  We would like to propose our method by considering unobserved events, namely counterfactuals, for counterfactual learning on missingnot-at-random feedback.  ... 
arXiv:2009.02623v2 fatcat:uo45zji6x5c6vdzy3swz5zzesq

Is it time for studying real-life debiasing? Evaluation of the effectiveness of an analogical intervention technique

Balazs Aczel, Bence Bago, Aba Szollosi, Andrei Foldes, Bence Lukacs
2015 Frontiers in Psychology  
For the purpose of the study, we devised an analogical debiasing technique for 10 biases (covariation detection, insensitivity to sample size, base rate neglect, regression to the mean, outcome bias, sunk  ...  The advantage of this method is that it can foster the transfer from learning abstract principles to improving behavioral performance.  ...  However, the transfer of this learning to new domains is rather elusive (Fong and Nisbett, 1991) .  ... 
doi:10.3389/fpsyg.2015.01120 pmid:26300816 pmcid:PMC4523707 fatcat:yxxkfr3kk5grzfamb3ln35w26m

Bias and Debias in Recommender System: A Survey and Future Directions [article]

Jiawei Chen, Hande Dong, Xiang Wang, Fuli Feng, Meng Wang, Xiangnan He
2021 arXiv   pre-print
We then provide a taxonomy to position and organize the existing work on recommendation debiasing.  ...  While recent years have witnessed a rapid growth of research papers on recommender system (RS), most of the papers focus on inventing machine learning models to better fit user behavior data.  ...  For example, [10] reviews explainable recommendation, [11] reviews knowledgebased recommendation, [12] and [13] summarize the recommendation methods based on deep learning and reinforcement learning  ... 
arXiv:2010.03240v2 fatcat:6fticc3otndsra2whs5e4nrdpi

Cross-domain User Preference Learning for Cold-start Recommendation [article]

Huiling Zhou, Jie Liu, Zhikang Li, Jin Yu, Hongxia Yang
2021 arXiv   pre-print
Cross-domain cold-start recommendation is an increasingly emerging issue for recommender systems.  ...  Existing works mainly focus on solving either cross-domain user recommendation or cold-start content recommendation.  ...  Cross-domain recommendation in realworld may involve several channels [44] .  ... 
arXiv:2112.03667v1 fatcat:fvg2amg5qber7encbsxgosxgtu

Debiasing Overconfidence among Indonesian Undergraduate Students in the Biology Classroom: An Intervention Study of the KAAR Model

Ai Nurlaelasari Rusmana, Fenny Roshayanti, Minsu Ha
2020 Asia-Pacific Science Education  
Recommendations for future intervention studies to reduce overconfidence among students are discussed.  ...  Metacognitive ability is enormously important for improving students' learning performance. However, overconfidence bias may hinder students' metacognition abilities.  ...  In general, there are two approaches for overconfidence debiasing: intervention designed and experience based.  ... 
doi:10.1163/23641177-bja00001 fatcat:d7d5irythzbkbnaev2otdgu4ri

Don't Generate Me: Training Differentially Private Generative Models with Sinkhorn Divergence [article]

Tianshi Cao, Alex Bie, Arash Vahdat, Sanja Fidler, Karsten Kreis
2021 arXiv   pre-print
We propose DP-Sinkhorn, a novel optimal transport-based generative method for learning data distributions from private data with differential privacy.  ...  Although machine learning models trained on massive data have led to break-throughs in several areas, their deployment in privacy-sensitive domains remains limited due to restricted access to data.  ...  Empirical OT distances are calculated between the "cross" group and the real data, and between the "debiasing" group and the "cross group".  ... 
arXiv:2111.01177v2 fatcat:7b2otssq5nc4fe4knbiuyyasqa

A Survey on Bias and Fairness in Machine Learning [article]

Ninareh Mehrabi, Fred Morstatter, Nripsuta Saxena, Kristina Lerman, Aram Galstyan
2019 arXiv   pre-print
We have recently seen work in machine learning, natural language processing, and deep learning that addresses such challenges in different subdomains.  ...  In addition to that, we examined different domains and subdomains in AI showing what researchers have observed with regard to unfair outcomes in the state-of-the-art methods and how they have tried to  ...  Table 2 While this section is largely domain-specific, it can be useful to take a cross-domain view.  ... 
arXiv:1908.09635v2 fatcat:ez4h4rzzgrf7njnbmsp7qnj24a

Intra-Processing Methods for Debiasing Neural Networks [article]

Yash Savani, Colin White, Naveen Sundar Govindarajulu
2020 arXiv   pre-print
As deep learning models become tasked with more and more decisions that impact human lives, such as criminal recidivism, loan repayment, and face recognition for law enforcement, bias is becoming a growing  ...  Creating debiasing algorithms specifically for this fine-tuning use-case has largely been neglected.  ...  Introduction The last decade has seen a huge increase in applications of machine learning in a wide variety of domains such as credit scoring, fraud detection, hiring decisions, criminal recidivism, loan  ... 
arXiv:2006.08564v2 fatcat:pauvm5izljcmldstpfhlbb3gda

Say No to the Discrimination: Learning Fair Graph Neural Networks with Limited Sensitive Attribute Information [article]

Enyan Dai, Suhang Wang
2021 arXiv   pre-print
As a result, the applications of GNNs in sensitive domains such as crime rate prediction would be largely limited.  ...  Extensive experiments on real-world datasets also demonstrate the effectiveness of FairGNN in debiasing and keeping high accuracy.  ...  ACKNOWLEDGEMENTS This material is based upon work supported by, or in part by, the National Science Foundation (NSF) under grant IIS-1909702, IIS-1955851, and the Global Research Outreach program of Samsung  ... 
arXiv:2009.01454v5 fatcat:wb2bqq4khfcxfim4k7h2fxkpou
« Previous Showing results 1 — 15 out of 519 results