Gabriel Erion
YOU?
Author Swipe
View article: Peer Review #2 of "Model independent feature attributions: Shapley values that uncover non-linear dependencies (v0.1)"
Peer Review #2 of "Model independent feature attributions: Shapley values that uncover non-linear dependencies (v0.1)" Open
Shapley values have become increasingly popular in the machine learning literature, thanks to their attractive axiomatisation, flexibility, and uniqueness in satisfying certain notions of 'fairness'.The flexibility arises from the myriad p…
View article: Peer Review #2 of "Model independent feature attributions: Shapley values that uncover non-linear dependencies (v0.2)"
Peer Review #2 of "Model independent feature attributions: Shapley values that uncover non-linear dependencies (v0.2)" Open
Shapley values have become increasingly popular in the machine learning literature, thanks to their attractive axiomatisation, flexibility, and uniqueness in satisfying certain notions of 'fairness'.The flexibility arises from the myriad p…
View article: suinleelab/attributionpriors: Nature Machine Intelligence code
suinleelab/attributionpriors: Nature Machine Intelligence code Open
Code submitted as the final revision of the manuscript "Improving performance of deep learning models with axiomatic attribution priors and expected gradients" at Nature Machine Intelligence.
View article: CoAI: Cost-Aware Artificial Intelligence for Health Care
CoAI: Cost-Aware Artificial Intelligence for Health Care Open
The recent emergence of accurate artificial intelligence (AI) models for disease diagnosis raises the possibility that AI-based clinical decision support could substantially lower the workload of healthcare providers. However, for this to …
View article: An adversarial approach for the robust classification of pneumonia from chest radiographs
An adversarial approach for the robust classification of pneumonia from chest radiographs Open
While deep learning has shown promise in the domain of disease classification from medical images, models based on state-of-the-art convolutional neural network architectures often exhibit performance loss due to dataset shift. Models trai…
View article: Deep Transfer Learning for Physiological Signals.
Deep Transfer Learning for Physiological Signals. Open
Deep learning is increasingly common in healthcare, yet transfer learning for physiological signals (e.g., temperature, heart rate, etc.) is under-explored. Here, we present a straightforward, yet performant framework for transferring know…
View article: An Adversarial Approach for the Robust Classification of Pneumonia from\n Chest Radiographs
An Adversarial Approach for the Robust Classification of Pneumonia from\n Chest Radiographs Open
While deep learning has shown promise in the domain of disease classification\nfrom medical images, models based on state-of-the-art convolutional neural\nnetwork architectures often exhibit performance loss due to dataset shift.\nModels t…
View article: Improving performance of deep learning models with axiomatic attribution\n priors and expected gradients
Improving performance of deep learning models with axiomatic attribution\n priors and expected gradients Open
Recent research has demonstrated that feature attribution methods for deep\nnetworks can themselves be incorporated into training; these attribution priors\noptimize for a model whose attributions have certain desirable properties --\nmost…
View article: Explainable AI for Trees: From Local Explanations to Global Understanding
Explainable AI for Trees: From Local Explanations to Global Understanding Open
Tree-based machine learning models such as random forests, decision trees, and gradient boosted trees are the most popular non-linear predictive models used in practice today, yet comparatively little attention has been paid to explaining …
View article: Consistent Individualized Feature Attribution for Tree Ensembles
Consistent Individualized Feature Attribution for Tree Ensembles Open
Interpreting predictions from tree ensemble methods such as gradient boosting machines and random forests is important, yet feature attribution for trees is often heuristic and not individualized for each prediction. Here we show that popu…
View article: Anesthesiologist-level forecasting of hypoxemia with only SpO2 data using deep learning
Anesthesiologist-level forecasting of hypoxemia with only SpO2 data using deep learning Open
We use a deep learning model trained only on a patient's blood oxygenation data (measurable with an inexpensive fingertip sensor) to predict impending hypoxemia (low blood oxygen) more accurately than trained anesthesiologists with access …
View article: Viral Genetic Linkage Analysis in the Presence of Missing Data
Viral Genetic Linkage Analysis in the Presence of Missing Data Open
Analyses of viral genetic linkage can provide insight into HIV transmission dynamics and the impact of prevention interventions. For example, such analyses have the potential to determine whether recently-infected individuals have acquired…
View article: Addressing Missing Data in Viral Genetic Linkage Analysis Through Multiple Imputation and Subsampling-Based Likelihood Optimization
Addressing Missing Data in Viral Genetic Linkage Analysis Through Multiple Imputation and Subsampling-Based Likelihood Optimization Open
This thesis addresses the intersection of two important areas in epidemiology and statistics: genetic linkage analysis and missing data methods, respectively. Genetic linkage analysis is a promising method in viral epidemiology which invol…