Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

169,291 papers · 148 categories

Trend · papers per month

145290434579 · Jun 202019922001200920182026
48 results for health-care applications

New fair regression methods improve health care spending predictions for undercompensated groups.

problem Current risk adjustment formulas underpredict spending for specific health groups, leading to unfair compensation.
method Developed new fair regression methods by integrating fairness considerations into the objective function.
result New methods lead to significant improvements in fairness (98%) with minimal impact on overall fit (4%).

A self-attention model improves fraud detection in health care claims.

problem Fraud detection in health care claims with hierarchical data structures.
method Piecewise feed forward neural networks and self-attention neural networks.
result Self-attention model outperforms other models on a dataset of two million health care claims.

Study on privacy-preserving health care models that sacrifice accuracy for data protection.

problem Privacy-preserving models in health care neglect data from the tails, reducing accuracy for small groups.
method Used state-of-the-art differentially private learning methods for clinical prediction tasks.
result Privacy-preserving models in health care exhibit steep tradeoffs between privacy and utility, and disproportionately influence large demographic groups.

Bayesian nonparametric models identify subgroups in hospital stay data.

problem Complex inpatient utilization data with zero inflation, over-dispersion, and skewness.
method Fully Bayesian mixture model with nonparametric clustering.
result Distinct subgroups of patients with lung cancer identified, differing in hospital days, covariates, and covariate relationships.

Icebreaker tackles active information acquisition with low data, improving model performance.

problem Deploying machine learning models with minimal training data, where each feature acquisition is costly.
method Proposes a full Bayesian Deep Latent Gaussian Model (BELGAM) with novel inference methods.
result Significantly outperforms previous models in small data settings, improving test time performance.

EDDI efficiently finds high-value information with minimal cost.

problem Balancing decision quality with acquisition cost in dynamic information acquisition.
method EDDI uses a partial variational autoencoder (Partial VAE) and an acquisition function to maximize expected information gain.
result Cost reduction and improved decision quality in benchmarks and real-world applications.

The paper evaluates the importance of monotonicity in AI fairness across various fields.

problem Ensuring fairness in AI applications across criminology, education, health care, and finance.
method Theoretical reasoning, simulation, and extensive empirical analysis of monotonic neural additive models (MNAMs).
result Monotonicity is essential for fairness in AI ethics and society, especially in criminology, education, health care, and finance.

PIMKL uses pathway knowledge to improve sample classification and provide interpretable molecular signatures.

problem Lack of reliable molecular biomarkers and poor generalization in machine learning for health care.
method Pathway Induced Multiple Kernel Learning (PIMKL) using a mixture of pathway-induced kernels optimized via Multiple Kernel Learning.
result PIMKL provides interpretable molecular signatures and stable predictions.

This study compares community detection algorithms for delineating health service areas.

problem Delineating health service areas to improve health care services.
method Comparative analysis of community detection algorithms on hospital-patient discharge networks.
result Infomap algorithm produced the best delineation of health service areas.

Approximate dynamic programming (ADP) has proven itself in a wide range of applications spanning large-scale transportation problems, health care, revenue management, and energy systems. The design of effective ADP algorithms has many dimensions, but one crucial factor is the stepsize rule used to update a value functi…

2014-07-10abs ↗pdf ↗

Improved predictions for rare labels using neural networks and ontologies.

problem Long-tailed frequency distribution in multi-label prediction problems.
method Modified neural network output layer with a Bayesian network of sigmoids leveraging ontology relationships.
result Significant improvements in per-label AUROC and average precision for less common labels.

Survey reviews explainability in AI for healthcare, emphasizing trust and transparency.

problem Lack of transparency hinders AI adoption in healthcare.
method Comprehensive literature review to guide explainable AI design.
result Quantitative evaluation metrics are needed for some explainability properties.

This research improves infection prediction from symptoms across diverse study designs.

problem Predicting infection from symptom data is challenging due to varied study formats and contexts.
method Assesses transfer learning methods for improving prediction across different study types.
result It is possible to use data from one study to predict infection in another, with performance close to or better than using a single dataset.

Estimates causal effect of managed care plans on NYC Medicaid spending.

problem Generalizing causal estimates to a target population not well-represented by randomized studies.
method Conditional cross-design synthesis estimators combining randomized and observational data.
result Estimates causal effect of managed care plans on health care spending.

Study on stock market volatility and return dispersion during COVID-19.

problem Impact of COVID-19 on stock market volatility and return dispersion.
method Used Google index to proxy epidemic impact, modeled volatility, and analyzed influencing factors of log-return.
result Volatility significantly affected by epidemic and cross-sectional return dispersion, with positive coefficients.

New fair regression method improves fairness in chronic kidney disease classification.

problem Mitigating societal bias in health care for multiple groups.
method Penalized fair regression framework for multiple groups, with penalties for true positive rate disparity.
result Achieves fairness-accuracy frontier beyond existing methods in simulations and real-world data.

Study assesses health plan risk measures for Solvency Capital Requirement.

problem Assessing risk measures for health plans to meet Solvency Capital Requirement.
method Three-part regression model with three GLMs for claim counts, episode allocation, and severity.
result Reduction in regression models compared to traditional methods.

GP-HD uses genetic programming to generate personalized health models.

problem Creating accurate, personalized health models from large health data.
method Genetic Programming framework to generate parameterized dynamical systems models.
result GP-HD models perform similarly to models based on domain knowledge and outperform LSTM models.

Deep-AER system uses lightweight CNN models for EEG-based emotion recognition.

problem Challenging task of emotion recognition using EEG signals.
method Two-level ensemble of lightweight 1D-CNN models trained on a small dataset.
result Deep-AER achieved high accuracies (98.43% and 97.65%) for emotion detection.

Improved DNN calibration without sacrificing accuracy.

problem Poor calibration of over-parametrized DNNs in safety-critical applications.
method Decoupling feature extraction and classification layers, and applying Gaussian priors.
result Significant improvement in model calibration with minimal training cost.

Differential privacy in distributed learning reduces privacy risks.

problem Protecting privacy in machine learning applications with distributed data.
method Secure multi-party sum function and Gaussian mechanism for differential privacy in a distributed setting.
result Asymptotically optimal and practically efficient DP Bayesian inference with diminishing extra cost.

New algorithm detects changes quickly without knowing parameters, near optimally.

problem Quickest change detection with unknown parameters.
method Leverages theoretical asymptotic properties to derive a scalable approximate algorithm with near optimal performance.
result Detects changes in constant complexity with near optimal performance.

Two machine learning models detect anomalies in ER claims, saving up to 40% in improper payments.

problem Improper health insurance payments from fraud and upcoding.
method Two machine learning models: an upcoding model based on severity code distributions and a random forest model for claim sorting.
result Random forest model saved 12% to 40% in improper payments compared to a baseline approach.

Distributed training of deep neural networks over multiple agents.

problem Scarcity of labeled data and computational resources in training neural networks.
method Distributed learning of deep neural networks over multiple agents using semi-supervised learning.
result Performance similar to a regular neural network trained on a single machine.

Generative models assess quality on time-series data using ITS and FITD.

problem Lack of consensus for quality assessment of class-conditional generative models on time-series data.
method Introduced InceptionTime Score (ITS) and Frechet InceptionTime Distance (FITD) to evaluate generative models.
result ITS and FITD combined with TSTR can accurately assess generative model performance on time-series data.

Improved robustness in multi-modal sensor fusion with deep learning.

problem Inconsistency in fusion weights leading to poor performance under sensor failures.
method Proposes deep multi-modal sensor fusion architectures with fusion weight regularization and target learning.
result Proposed architectures outperform existing deep learning methods under sensor failures.

Paper develops deep learning for predicting clinical endpoints from diverse medical records.

problem Predicting clinical endpoints from heterogeneous, irregularly visiting medical records.
method Proposes a novel model with a new gate to control visiting rates of different events.
result Model effectively predicts death and abnormal lab tests with real-world clinical data.

Many applications collect a large number of time series, for example, the financial data of companies quoted in a stock exchange, the health care data of all patients that visit the emergency room of a hospital, or the temperature sequences continuously measured by weather stations across the US. These data are often r…

2015-02-28abs ↗pdf ↗

Alternative model predicts health insurance reimbursement based on contract limitations.

problem Estimating the ratio of reimbursement to health care expenditures after deductibles and copayments.
method Proposes a Zero-One Inflated Beta regression model using GAMLSS.
result The model provides a dependency structure between reimbursement and contract limitations.