Using observational data to learn causal relationships is essential when randomized experiments are not possible, such as in healthcare. Discovering causal relationships in time-series health data is even more challenging when relationships change over the course of a disease, such as medications that are most effective early on or for individuals with severe disease. Stage variables such as weeks of pregnancy, disease stages, or biomarkers like HbA1c, can influence what causal relationships are true for a patient. However, causal inference within each stage is often not possible due to limited amounts of data, and combining all data risks incorrect or missed inferences. To address this, we propose Causal Discovery with Stage Variables (CDSV), which uses stage variables to reweight data from multiple time-series while accounting for different causal relationships in each stage. In simulated data, CDSV discovers more causes with fewer false discoveries compared to baselines, in eICU it has a lower FDR than baselines, and in MIMIC-III it discovers more clinically relevant causes of high blood pressure.

We propose a novel graph-regularized neural network (GRNN) algorithm for tree species classification. The proposed algorithm encompasses superpixel-based segmentation for graph construction, a pixel-wise neural network classifier, and the label propagation technique to generate an accurate and realistic (emulating tree crowns) classification map on a sparsely annotated data set. GRNN outperforms several state-of-the-art techniques not only for the standard Indian Pines HSI but also achieves a high classification accuracy (approx. 92%) on a new HSI data set collected over the heterogeneous forests of French Guiana (FG) when less than 1% of the pixels are labeled. We further show that GRNN is competitive with the state-of-the-art semi-supervised methods and exhibits a small deviation in accuracy for different numbers of training samples and over repeated trials with randomly sampled labeled pixels for training.

With the aim to stimulate future research, we describe an exploratory study of a railway rescheduling problem. A widely used approach in practice and state of the art is to decompose these complex problems by geographical scope. Instead, we propose defining a core problem that restricts a rescheduling problem in response to a disturbance to only trains that need to be rescheduled, hence restricting the scope in both time and space. In this context, the difficulty resides in defining a scoper that can predict a subset of train services that will be affected by a given disturbance. We report preliminary results using the Flatland simulation environment that highlights the potential and challenges of this idea. We provide an extensible playground open-source implementation based on the Flatland railway environment and Answer-Set Programming.

DFU is a severe complication of diabetes that can lead to amputation of the lower limb if not treated properly. Inspired by the 2021 Diabetic Foot Ulcer Grand Challenge, researchers designed automated multi-class classification of DFU, including infection, ischaemia, both of these conditions, and none of these conditions. However, it remains a challenge as classification accuracy is still not satisfactory. This paper proposes a Venn Diagram interpretation of multi-label CNN-based method, utilizing different image enhancement strategies, to improve the multi-class DFU classification. We propose to reduce the four classes into two since both class wounds can be interpreted as the simultaneous occurrence of infection and ischaemia and none class wounds as the absence of infection and ischaemia. We introduce a novel Venn Diagram representation block in the classifier to interpret all four classes from these two classes. To make our model more resilient, we propose enhancing the perceptual quality of DFU images, particularly blurry or inconsistently lit DFU images, by performing color and sharpness enhancements on them. We also employ a fine-tuned optimization technique, adaptive sharpness aware minimization, to improve the CNN model generalization performance. The proposed method is evaluated on the test dataset of DFUC2021, containing 5,734 images and the results are compared with the top-3 winning entries of DFUC2021. Our proposed approach outperforms these existing approaches and achieves Macro-Average F1, Recall and Precision scores of 0.6592, 0.6593, and 0.6652, respectively.Additionally, We perform ablation studies and image quality measurements to further interpret our proposed method. This proposed method will benefit patients with DFUs since it tackles the inconsistencies in captured images and can be employed for a more robust remote DFU wound classification.

Federated Learning (FL) is a novel approach enabling several clients holding sensitive data to collaboratively train machine learning models, without centralizing data. The cross-silo FL setting corresponds to the case of few ($2$--$50$) reliable clients, each holding medium to large datasets, and is typically found in applications such as healthcare, finance, or industry. While previous works have proposed representative datasets for cross-device FL, few realistic healthcare cross-silo FL datasets exist, thereby slowing algorithmic research in this critical application. In this work, we propose a novel cross-silo dataset suite focused on healthcare, FLamby (Federated Learning AMple Benchmark of Your cross-silo strategies), to bridge the gap between theory and practice of cross-silo FL. FLamby encompasses 7 healthcare datasets with natural splits, covering multiple tasks, modalities, and data volumes, each accompanied with baseline training code. As an illustration, we additionally benchmark standard FL algorithms on all datasets. Our flexible and modular suite allows researchers to easily download datasets, reproduce results and re-use the different components for their research. FLamby is available at~\url{www.github.com/owkin/flamby}.

Most existing exoplanets are discovered using validation techniques rather than being confirmed by complementary observations. These techniques generate a score that is typically the probability of the transit signal being an exoplanet (y(x)=exoplanet) given some information related to that signal (represented by x). Except for the validation technique in Rowe et al. (2014) that uses multiplicity information to generate these probability scores, the existing validation techniques ignore the multiplicity boost information. In this work, we introduce a framework with the following premise: given an existing transit signal vetter (classifier), improve its performance using multiplicity information. We apply this framework to several existing classifiers, which include vespa (Morton et al. 2016), Robovetter (Coughlin et al. 2017), AstroNet (Shallue & Vanderburg 2018), ExoNet (Ansdel et al. 2018), GPC and RFC (Armstrong et al. 2020), and ExoMiner (Valizadegan et al. 2022), to support our claim that this framework is able to improve the performance of a given classifier. We then use the proposed multiplicity boost framework for ExoMiner V1.2, which addresses some of the shortcomings of the original ExoMiner classifier (Valizadegan et al. 2022), and validate 69 new exoplanets for systems with multiple KOIs from the Kepler catalog.

Machine learning (ML) approaches have been used to develop highly accurate and efficient applications in many fields including bio-medical science. However, even with advanced ML techniques, cancer classification using gene expression data is still complicated because of the high dimensionality of the datasets employed. We developed a new fuzzy gene selection technique (FGS) to identify informative genes to facilitate cancer classification and reduce the dimensionality of the available gene expression data. Three feature selection methods (Mutual Information, F-ClassIf, and Chi-squared) were evaluated and employed to obtain the score and rank for each gene. Then, using Fuzzification and Defuzzification methods to obtain the best single score for each gene, which aids in the identification of significant genes. Our study applied the fuzzy measures to six gene expression datasets including four Microarray and two RNA-seq datasets for evaluating the proposed algorithm. With our FGS-enhanced method, the cancer classification model achieved 96.5%,96.2%,96%, and 95.9% for accuracy, precision, recall, and f1-score respectively, which is significantly higher than 69.2% accuracy, 57.8% precision, 66% recall, and 58.2% f1-score when the standard MLP method was used. In examining the six datasets that were used, the proposed model demonstrates it's capacity to classify cancer effectively.

Monitoring the health status of patients in the Intensive Care Unit (ICU) is a critical aspect of providing superior care and treatment. The availability of large-scale electronic health records (EHR) provides machine learning models with an abundance of clinical text and vital sign data, enabling them to make highly accurate predictions. Despite the emergence of advanced Natural Language Processing (NLP) algorithms for clinical note analysis, the complex textual structure and noise present in raw clinical data have posed significant challenges. Coarse embedding approaches without domain-specific refinement have limited the accuracy of these algorithms. To address this issue, we propose FINEEHR, a system that utilizes two representation learning techniques, namely metric learning and fine-tuning, to refine clinical note embeddings, while leveraging the intrinsic correlations among different health statuses and note categories. We evaluate the performance of FINEEHR using two metrics, namely Area Under the Curve (AUC) and AUC-PR, on a real-world MIMIC III dataset. Our experimental results demonstrate that both refinement approaches improve prediction accuracy, and their combination yields the best results. Moreover, our proposed method outperforms prior works, with an AUC improvement of over 10%, achieving an average AUC of 96.04% and an average AUC-PR of 96.48% across various classifiers.

While transformer-based systems have enabled greater accuracies with fewer training examples, data acquisition obstacles still persist for rare-class tasks -- when the class label is very infrequent (e.g. < 5% of samples). Active learning has in general been proposed to alleviate such challenges, but choice of selection strategy, the criteria by which rare-class examples are chosen, has not been systematically evaluated. Further, transformers enable iterative transfer-learning approaches. We propose and investigate transfer- and active learning solutions to the rare class problem of dissonance detection through utilizing models trained on closely related tasks and the evaluation of acquisition strategies, including a proposed probability-of-rare-class (PRC) approach. We perform these experiments for a specific rare class problem: collecting language samples of cognitive dissonance from social media. We find that PRC is a simple and effective strategy to guide annotations and ultimately improve model accuracy while transfer-learning in a specific order can improve the cold-start performance of the learner but does not benefit iterations of active learning.

Selecting a minimal feature set that is maximally informative about a target variable is a central task in machine learning and statistics. Information theory provides a powerful framework for formulating feature selection algorithms -- yet, a rigorous, information-theoretic definition of feature relevancy, which accounts for feature interactions such as redundant and synergistic contributions, is still missing. We argue that this lack is inherent to classical information theory which does not provide measures to decompose the information a set of variables provides about a target into unique, redundant, and synergistic contributions. Such a decomposition has been introduced only recently by the partial information decomposition (PID) framework. Using PID, we clarify why feature selection is a conceptually difficult problem when approached using information theory and provide a novel definition of feature relevancy and redundancy in PID terms. From this definition, we show that the conditional mutual information (CMI) maximizes relevancy while minimizing redundancy and propose an iterative, CMI-based algorithm for practical feature selection. We demonstrate the power of our CMI-based algorithm in comparison to the unconditional mutual information on benchmark examples and provide corresponding PID estimates to highlight how PID allows to quantify information contribution of features and their interactions in feature-selection problems.