In this post, I will present how to use random forests in classification, a prediction technique consisting in generating a set of trees (hence, a forest) bootstrapping the features used in each tree. We do this to obtain trees that are not necessarily using the strongest predictors at the beginning. I will test this technique in a LoanDefaults dataset to predict which customers will default the paying of a loan in a specific month. This dataset has two interesting features: the number of positive cases is much smaller than the negatives and requires some preprocessing of the existing features. I will be using the ranger (RANdom forest GEneRator) package, skimr to get a summary of data, rpart and rpart.plot to generate an alternative decision tree model, BAdatasets to access the dataset, tidymodels for prediction workflow facilities and forcats for the variable importance plot.

Molecular property prediction is of crucial importance in many disciplines such as drug discovery, molecular biology, or material and process design. The frequently employed quantitative structure-property/activity relationships (QSPRs/QSARs) characterize molecules by descriptors which are then mapped to the properties of interest via a linear or nonlinear model. In contrast, graph neural networks, a novel machine learning method, directly work on the molecular graph, i.e., a graph representation where atoms correspond to nodes and bonds correspond to edges. GNNs allow to learn properties in an end-to-end fashion, thereby avoiding the need for informative descriptors as in QSPRs/QSARs. GNNs have been shown to achieve state-of-the-art prediction performance on various property predictions tasks and represent an active field of research. We describe the fundamentals of GNNs and demonstrate the application of GNNs via two examples for molecular property prediction.

Plasma is defined as the fourth state of matter and non-thermal plasma can be produced at atmospheric pressure under a high electrical field. The strong and broad-spectrum antimicrobial effect of plasma-activated liquids (PALs) is now well known. The proven applicability of machine learning (ML) in the medical field is encouraging for its application in the field of plasma medicine as well. Thus, ML applications on PALs could present a new perspective to better understand the influences of various parameters on their antimicrobial effects. In this paper, comparative supervised ML models are presented by using previously obtained data to qualitatively predict the in vitro antimicrobial activity of PALs. A literature search was performed and data is collected from 33 relevant articles. After the required preprocessing steps, two supervised ML methods, namely classification, and regression are applied to data to obtain microbial inactivation (MI) predictions. For classification, MI is labeled in four categories and for regression, MI is used as a continuous variable. Two different robust cross-validation strategies are conducted for classification and regression models to evaluate the proposed method; repeated stratified k-fold cross-validation and k-fold cross-validation, respectively. We also investigate the effect of different features on models. The results demonstrated that the hyperparameter-optimized Random Forest Classifier (oRFC) and Random Forest Regressor (oRFR) provided better results than other models for the classification and regression, respectively. Finally, the best test accuracy of 82.68% for oRFC and R2 of 0.75 for the oRFR are obtained. ML techniques could contribute to a better understanding of plasma parameters that have a dominant role in the desired antimicrobial effect. Furthermore, such findings may contribute to the definition of a plasma dose in the future.

Self-supervised, category-agnostic segmentation of real-world images is a challenging open problem in computer vision. Here, we show how to learn static grouping priors from motion self-supervision by building on the cognitive science concept of a Spelke Object: a set of physical stuff that moves together. We introduce the Excitatory-Inhibitory Segment Extraction Network (EISEN), which learns to extract pairwise affinity graphs for static scenes from motion-based training signals. EISEN then produces segments from affinities using a novel graph propagation and competition network. During training, objects that undergo correlated motion (such as robot arms and the objects they move) are decoupled by a bootstrapping process: EISEN explains away the motion of objects it has already learned to segment. We show that EISEN achieves a substantial improvement in the state of the art for self-supervised image segmentation on challenging synthetic and real-world robotics datasets.

Integrated Development Environments (IDEs) provide tool support to automate many source code editing tasks. Traditionally, IDEs use only the spatial context, i.e., the location where the developer is editing, to generate candidate edit recommendations. However, spatial context alone is often not sufficient to confidently predict the developer's next edit, and thus IDEs generate many suggestions at a location. Therefore, IDEs generally do not actively offer suggestions and instead, the developer is usually required to click on a specific icon or menu and then select from a large list of potential suggestions. As a consequence, developers often miss the opportunity to use the tool support because they are not aware it exists or forget to use it. To better understand common patterns in developer behavior and produce better edit recommendations, we can additionally use the temporal context, i.e., the edits that a developer was recently performing. To enable edit recommendations based on temporal context, we present Overwatch, a novel technique for learning edit sequence patterns from traces of developers' edits performed in an IDE. Our experiments show that Overwatch has 78% precision and that Overwatch not only completed edits when developers missed the opportunity to use the IDE tool support but also predicted new edits that have no tool support in the IDE.

Human-in-the-loop approaches are of great importance for robot applications. In the presented study, we implemented a multimodal human-robot interaction (HRI) scenario, in which a simulated robot communicates with its human partner through speech and gestures. The robot announces its intention verbally and selects the appropriate action using pointing gestures. The human partner, in turn, evaluates whether the robot's verbal announcement (intention) matches the action (pointing gesture) chosen by the robot. For cases where the verbal announcement of the robot does not match the corresponding action choice of the robot, we expect error-related potentials (ErrPs) in the human electroencephalogram (EEG). These intrinsic evaluations of robot actions by humans, evident in the EEG, were recorded in real time, continuously segmented online and classified asynchronously. For feature selection, we propose an approach that allows the combinations of forward and backward sliding windows to train a classifier. We achieved an average classification performance of 91% across 9 subjects. As expected, we also observed a relatively high variability between the subjects. In the future, the proposed feature selection approach will be extended to allow for customization of feature selection. To this end, the best combinations of forward and backward sliding windows will be automatically selected to account for inter-subject variability in classification performance. In addition, we plan to use the intrinsic human error evaluation evident in the error case by the ErrP in interactive reinforcement learning to improve multimodal human-robot interaction.

The lack of well-calibrated confidence estimates makes neural networks inadequate in safety-critical domains such as autonomous driving or healthcare. In these settings, having the ability to abstain from making a prediction on out-of-distribution (OOD) data can be as important as correctly classifying in-distribution data. We introduce $p$-DkNN, a novel inference procedure that takes a trained deep neural network and analyzes the similarity structures of its intermediate hidden representations to compute $p$-values associated with the end-to-end model prediction. The intuition is that statistical tests performed on latent representations can serve not only as a classifier, but also offer a statistically well-founded estimation of uncertainty. $p$-DkNN is scalable and leverages the composition of representations learned by hidden layers, which makes deep representation learning successful. Our theoretical analysis builds on Neyman-Pearson classification and connects it to recent advances in selective classification (reject option). We demonstrate advantageous trade-offs between abstaining from predicting on OOD inputs and maintaining high accuracy on in-distribution inputs. We find that $p$-DkNN forces adaptive attackers crafting adversarial examples, a form of worst-case OOD inputs, to introduce semantically meaningful changes to the inputs.

The approval success rate of drug candidates is very low with the majority of failure due to safety and efficacy. Increasingly available high dimensional information on targets, drug molecules and indications provides an opportunity for ML methods to integrate multiple data modalities and better predict clinically promising drug targets. Notably, drug targets with human genetics evidence are shown to have better odds to succeed. However, a recent tensor factorization-based approach found that additional information on targets and indications might not necessarily improve the predictive accuracy. Here we revisit this approach by integrating different types of human genetics evidence collated from publicly available sources to support each target-indication pair. We use Bayesian tensor factorization to show that models incorporating all available human genetics evidence (rare disease, gene burden, common disease) modestly improves the clinical outcome prediction over models using single line of genetics evidence. We provide additional insight into the relative predictive power of different types of human genetics evidence for predicting the success of clinical outcomes.

This paper explores the effectiveness of data-driven models to predict voltage excursion events in power systems using simple categorical labels. By treating the prediction as a categorical classification task, the workflow is characterized by a low computational and data burden. A proof-of-concept case study on a real portion of the Italian 150 kV sub-transmission network, which hosts a significant amount of wind power generation, demonstrates the general validity of the proposal and offers insight into the strengths and weaknesses of several widely utilized prediction models for this application.

Neuroimaging studies are often limited by the number of subjects and cognitive processes that can be feasibly interrogated. However, a rapidly growing number of neuroscientific studies have collectively accumulated an extensive wealth of results. Digesting this growing literature and obtaining novel insights remains to be a major challenge, since existing meta-analytic tools are constrained to keyword queries. In this paper, we present Text2Brain, an easy to use tool for synthesizing brain activation maps from open-ended text queries. Text2Brain was built on a transformer-based neural network language model and a coordinate-based meta-analysis of neuroimaging studies. Text2Brain combines a transformer-based text encoder and a 3D image generator, and was trained on variable-length text snippets and their corresponding activation maps sampled from 13,000 published studies. In our experiments, we demonstrate that Text2Brain can synthesize meaningful neural activation patterns from various free-form textual descriptions. Text2Brain is available at https://braininterpreter.com as a web-based tool for efficiently searching through the vast neuroimaging literature and generating new hypotheses.