The use of machine learning techniques in biomedical research has exploded over the past few years, as exemplified by the dramatic increase in the number of journal articles indexed on PubMed by the term "machine learning", from 3,200 in 2015 to over 18,000 in 2020. While substantial scientific advancements have been made possible thanks to machine learning, the inner working of most machine learning algorithms remains foreign to many clinicians, most of whom are quite familiar with traditional statistical methods but have little formal training on advanced computer algorithms. Unfortunately, journal reviewers and editors are sometimes content with accepting machine learning as a black box operation and fail to analyze the results produced by machine learning models with the same level of scrutiny that is applied to other clinical and basic science research. The goal of this journal club is to help readers develop the knowledge and skills necessary to digest and critique biomedical journal articles involving the use of machine learning techniques. It is hard for a reviewer to know what questions to ask if he/she does not understand how these algorithms work.

Originally published on Towards AI the World's Leading AI and Technology News and Media Company. If you are building an AI-related product or service, we invite you to consider becoming an AI sponsor. At Towards AI, we help scale AI and technology startups. Let us help you unleash your technology to the masses. When it comes to articles on deep learning, advances in Computer Vision or Natural Language Processing (NLP) receive the lion's share of the attention.

"Just Accepted" papers have undergone full peer review and have been accepted for publication in Radiology: Artificial Intelligence. This article will undergo copyediting, layout, and proof review before it is published in its final version. Please note that during production of the final copyedited article, errors may be discovered which could affect the content. To assess if semisupervised natural language processing (NLP) of text clinical radiology reports could provide useful automated diagnosis categorization for ground truth labeling to overcome manual labeling bottlenecks in the machine learning pipeline. In this retrospective study, 1503 text cardiac MRI reports (from between 2016 and 2019) were manually annotated for five diagnoses by clinicians: normal, dilated cardiomyopathy (DCM), hypertrophic cardiomyopathy (HCM), myocardial infarction (MI), and myocarditis.

Now, we get into the more technical part of this article: ML Algorithms. I'll include a TL;DR, although I encourage you to browse through the different algorithms. I'll also include a really helpful AI course video that I personally used to learn and write this article at the very end. Without further wait, let's start going in-depth on the different algorithms used in machine learning. Linear regression is a supervised learning model used for regression that predicts a dependent variable (Y) based on the independent variable(s) (X) by fitting a line linearly.

Energy-based modeling is a promising approach to unsupervised learning, which yields many downstream applications from a single model. The main difficulty in learning energy-based models with the "contrastive approaches" is the generation of samples from the current energy function at each iteration. Many advances have been made to accomplish this subroutine cheaply. Nevertheless, all such sampling paradigms run MCMC targeting the current model, which requires infinitely long chains to generate samples from the true energy distribution and is problematic in practice. This paper proposes an alternative approach to getting these samples and avoiding crude MCMC sampling from the current model. We accomplish this by viewing the evolution of the modeling distribution as (i) the evolution of the energy function, and (ii) the evolution of the samples from this distribution along some vector field. We subsequently derive this time-dependent vector field such that the particles following this field are approximately distributed as the current density model. Thereby we match the evolution of the particles with the evolution of the energy function prescribed by the learning procedure. Importantly, unlike Monte Carlo sampling, our method targets to match the current distribution in a finite time. Finally, we demonstrate its effectiveness empirically compared to MCMC-based learning methods.

This paper addresses a problem called "conditional manifold learning", which aims to learn a low-dimensional manifold embedding of high-dimensional data, conditioning on auxiliary manifold information. This auxiliary manifold information is from controllable or measurable conditions, which are ubiquitous in many science and engineering applications. A broad class of solutions for this problem, conditional multidimensional scaling (including a conditional ISOMAP variant), is proposed. A conditional version of the SMACOF algorithm is introduced to optimize the objective function of conditional multidimensional scaling.

We establish an equivalence between a family of adversarial training problems for non-parametric binary classification and a family of regularized risk minimization problems where the regularizer is a nonlocal perimeter functional. The resulting regularized risk minimization problems admit exact convex relaxations of the type $L^1+$ (nonlocal) $\operatorname{TV}$, a form frequently studied in image analysis and graph-based learning. A rich geometric structure is revealed by this reformulation which in turn allows us to establish a series of properties of optimal solutions of the original problem, including the existence of minimal and maximal solutions (interpreted in a suitable sense), and the existence of regular solutions (also interpreted in a suitable sense). In addition, we highlight how the connection between adversarial training and perimeter minimization problems provides a novel, directly interpretable, statistical motivation for a family of regularized risk minimization problems involving perimeter/total variation. The majority of our theoretical results are independent of the distance used to define adversarial attacks.

Explainable artificial intelligence (XAI) and interpretable machine learning (IML) have become active research fields in recent years (Adadi and Berrada 2018; Molnar 2019). This is a natural consequence as complex machine learning (ML) models are now applied to solve supervised learning problems in many high-risk areas: cancer prognosis (Kourou et al. 2015), credit scoring (Kvamme et al. 2018), and money laundering detection (Jullum, Løland, et al. 2020). The high prediction accuracy of complex ML models often comes at the expense of model interpretability. As the goal of science is to gain knowledge from the collected data, the use of black-box models hinders the understanding of the underlying relationship between the features and the response, and thereby curtail scientific discovery. Model explanation frameworks from the XAI field extract the hidden knowledge about the underlying data structure captured by a black-box model, and thereby make the model's decision-making process transparent. This is crucial for, e.g., medical researchers that apply an ML model to obtain well-performing predictions, but who simultaneously also strive to discover important risk factors. Another driving factor is the Right to Explanation legislation in EU's General Data Protection Regulation (GDPR) (European Commission 2016).

Offline policy learning (OPL) leverages existing data collected a priori for policy optimization without any active exploration. Despite the prevalence and recent interest in this problem, its theoretical and algorithmic foundations in function approximation settings remain under-developed. In this paper, we consider this problem on the axes of distributional shift, optimization, and generalization in offline contextual bandits with neural networks. In particular, we propose a provably efficient offline contextual bandit with neural network function approximation that does not require any functional assumption on the reward. We show that our method provably generalizes over unseen contexts under a milder condition for distributional shift than the existing OPL works. Notably, unlike any other OPL method, our method learns from the offline data in an online manner using stochastic gradient descent, allowing us to leverage the benefits of online learning into an offline setting. Moreover, we show that our method is more computationally efficient and has a better dependence on the effective dimension of the neural network than an online counterpart. Finally, we demonstrate the empirical effectiveness of our method in a range of synthetic and real-world OPL problems.

Feature selection is an extensively studied technique in the machine learning literature where the main objective is to identify the subset of features that provides the highest predictive power. However, in causal inference, our goal is to identify the set of variables that are associated with both the treatment variable and outcome (i.e., the confounders). While controlling for the confounding variables helps us to achieve an unbiased estimate of causal effect, recent research shows that controlling for purely outcome predictors along with the confounders can reduce the variance of the estimate. In this paper, we propose an Outcome Adaptive Elastic-Net (OAENet) method specifically designed for causal inference to select the confounders and outcome predictors for inclusion in the propensity score model or in the matching mechanism. OAENet provides two major advantages over existing methods: it performs superiorly on correlated data, and it can be applied to any matching method and any estimates. In addition, OAENet is computationally efficient compared to state-of-the-art methods.