Colleges and universities are increasingly turning to algorithms that predict college-student success inform various decisions, including those related admissions, budgeting, student-success interventions. Because predictive rely on historical data, they capture societal injustices, racism. In this study, we examine how the accuracy of college student predictions differs between racialized groups, signaling algorithmic bias. We also evaluate utility leading bias-mitigating techniques in...

The education sector has been quick to recognize the power of predictive analytics enhance student success rates. However, there are challenges widespread adoption, including lack accessibility and potential perpetuation inequalities. These present in different stages modeling, data preparation, model development, evaluation. steps can introduce additional bias system if not appropriately performed. Substantial incompleteness responses is a common problem nationally representative at large...

Abstract Background Intensive Care Unit (ICU) readmissions in patients with heart failure (HF) result a significant risk of death and financial burden for healthcare systems. Prediction at-risk readmission allows targeted interventions that reduce morbidity mortality. Methods results We presented process mining/deep learning approach the prediction unplanned 30-day ICU HF. A patient’s health records can be understood as sequence observations called event logs; used to discover model. Time...

The exploration-exploitation trade-off poses a significant challenge in surrogate optimization for expensive black-box functions, particularly when dealing with batch evaluation settings. Despite efforts to develop sampling techniques, they often fall short of sufficiently prioritizing diversity within the selected batch. In this paper, we propose fundamentally novel approach called DPP-based Surrogate Optimization (DPPSO), which serves as consolidated framework. DPPSO introduces...

Machine learning (ML) is increasingly being used in high-stakes applications impacting society. Therefore, it of critical importance that ML models do not propagate discrimination. Collecting accurate labeled data societal applications challenging and costly. Active a promising approach to build an classifier by interactively querying oracle within a labeling budget. We design algorithms for fair active carefully selects points to be so as balance model accuracy fairness. We demonstrate the...

Explaining complex algorithms and models has recently received growing attention in various domains to support informed decisions. Ranking functions are widely used for almost every form of human activity enable effective decision-making processes. Hence, explaining ranking indicators their importance essential properties enhance performance. Local explanation techniques have become a prominent way interpret individual predictions machine learning models. However, there been limited...

Abstract The cetane number (CN) is an important fuel property to consider for compression ignition engines as it a measure of fuel's delay. Derived (DCN) already varies significantly within jet fuels. With the expected increasing prevalence alternative fuels, additional variability expected. DCN usually assigned fuels using ASTM methods that use large equipment like quality tester (IQT), which consumes lot and cumbersome operate. Over last decade, there have been advances in development...

As agencies consider strategies for meeting the livability goals (e.g., public health, mobility, and access) of communities, they often repurpose existing right-of-way arterials to meet traffic operation or safety goals. Agencies need more definitive guidance on when this repurposing can be done without negatively affecting safety. Researchers around world have considered efficacy lane width, but none defines width design recommendations urban arterials. This study uses treed regression...

Predictive analytics is widely used in various domains, including education, to inform decision-making and improve outcomes. However, many predictive models are proprietary inaccessible for evaluation or modification by researchers practitioners, limiting their accountability ethical design. Moreover, often opaque incomprehensible the officials who use them, reducing trust utility. Furthermore, may introduce exacerbate bias inequity, as they have done sectors of society. Therefore, there a...

Predictive analytics has been widely used in various domains, including education, to inform decision-making and improve outcomes. However, many predictive models are proprietary inaccessible for evaluation or modification by researchers practitioners, limiting their accountability ethical design. Moreover, often opaque incomprehensible the officials who use them, reducing trust utility. Furthermore, may introduce exacerbate bias inequity, as they have done sectors of society. Therefore,...

Optimizing costly black-box functions within a constrained evaluation budget presents significant challenges in many real-world applications. Surrogate Optimization (SO) is common resolution, yet its proprietary nature introduced by the complexity of surrogate models and sampling core (e.g., acquisition functions) often leads to lack explainability transparency. While existing literature has primarily concentrated on enhancing convergence global optima, practical interpretation newly...

Despite the potential benefits of machine learning (ML) in high-risk decision-making domains, deployment ML is not accessible to practitioners, and there a risk discrimination. To establish trust acceptance such democratizing tools fairness consideration are crucial. In this paper, we introduce FairPilot, an interactive system designed promote responsible development models by exploring combination various models, different hyperparameters, wide range definitions. We emphasize challenge...

Abstract The cetane number is an important fuel property to consider for compression ignition engines as it a measure of fuel’s delay. Derived (DCN) already varies significantly within jet fuels. With the expected increasing prevalence alternative fuels, additional variability expected. DCN usually assigned fuels using ASTM methods that use large equipment like Ignition Quality Tester (IQT), which consumes lot and cumbersome operate. Over last decade, there have been advances in development...

Safety problems of high energy density and high-power Li-ion batteries (LIBs) associated with thermal runaway (TRA) have become a serious problem resulting in massive recalls by manufacturers, at times even endangering the lives consumers. In general, TRA consists battery’s rapid self-heating sourced from exothermic (electro-)chemical reactions and/or mechanical abuse. It is impossible to directly monitor occurring events during practical operating conditions (e.g., cell phones or electric...

Machine learning (ML) is increasingly being used in high-stakes applications impacting society. Therefore, it of critical importance that ML models do not propagate discrimination. Collecting accurate labeled data societal challenging and costly. Active a promising approach to build an classifier by interactively querying oracle within labeling budget. We design algorithms for fair active carefully selects points be so as balance model accuracy fairness. demonstrate the effectiveness...

This paper reviews the state-of-the-art model-based adaptive sampling approaches for single-objective black-box optimization (BBO). While BBO literature includes various promising techniques, there is still a lack of comprehensive investigations existing research across vast scope problems. We first classify problems into two categories engineering design and algorithm discuss their challenges. then critically analyze techniques focusing on key acquisition functions. elaborate shortcomings...

Colleges and universities are increasingly turning to algorithms that predict college-student success inform various decisions, including those related admissions, budgeting, student-success interventions. Because predictive rely on historical data, they capture societal injustices, racism. A model includes racial categories may racially minoritized students will have less favorable outcomes. In this study, we explore bias in education data by modeling bachelor's degree attainment using...