Self-supervised pretraining has been observed to improve performance in supervised learning tasks in medical imaging. This study investigates the utility of self-supervised pretraining prior to conducting supervised fine-tuning for the downstream task of lung sliding classification in M-mode lung ultrasound images. We propose a novel pairwise relationship that couples M-mode images constructed from the same B-mode image and investigate the utility of data augmentation procedure specific to M-mode lung ultrasound. The results indicate that self-supervised pretraining yields better performance than full supervision, most notably for feature extractors not initialized with ImageNet-pretrained weights. Moreover, we observe that including a vast volume of unlabelled data results in improved performance on external validation datasets, underscoring the value of self-supervision for improving generalizability in automatic ultrasound interpretation. To the authors' best knowledge, this study is the first to characterize the influence of self-supervised pretraining for M-mode ultrasound.

Polypharmacy, most often defined as the simultaneous consumption of five or more drugs at once, is a prevalent phenomenon in the older population. Some of these polypharmacies, deemed inappropriate, may be associated with adverse health outcomes such as death or hospitalization. Considering the combinatorial nature of the problem as well as the size of claims database and the cost to compute an exact association measure for a given drug combination, it is impossible to investigate every possible combination of drugs. Therefore, we propose to optimize the search for potentially inappropriate polypharmacies (PIPs). To this end, we propose the OptimNeuralTS strategy, based on Neural Thompson Sampling and differential evolution, to efficiently mine claims datasets and build a predictive model of the association between drug combinations and health outcomes. We benchmark our method using two datasets generated by an internally developed simulator of polypharmacy data containing 500 drugs and 100 000 distinct combinations. Empirically, our method can detect up to 72% of PIPs while maintaining an average precision score of 99% using 30 000 time steps.

In recent years, neural networks of various configurations have been increasingly used to analyze data obtained with fluorescent and Cherenkov telescopes. In particular, a whole series of studies dedicated to the analysis of gamma-ray astronomy data with neural networks has been performed by the VERITAS [1], TAIGA [2, 3], and CTA [4, 5] collaborations. Typical tasks are the recognition of particular signal patterns in the data flow. In the simplest case, the problem can be reduced to classifying data into two groups: data samples that contain a signal of the desired type and all the rest. Since data obtained with the help of telescopes can naturally be considered as images or animations, one of the popular tools for classifying them are convolutional neural networks (CNNs), created primarily for image classification. CNNs have demonstrated the highest efficiency in this class of problems, see, for example, [6, 7].

The deployment of machine learning classifiers in high-stakes domains requires well-calibrated confidence scores for model predictions. In this paper we introduce the notion of variable-based calibration to characterize calibration properties of a model with respect to a variable of interest, generalizing traditional score-based metrics such as expected calibration error (ECE). In particular, we find that models with near-perfect ECE can exhibit significant miscalibration as a function of features of the data. We demonstrate this phenomenon both theoretically and in practice on multiple well-known datasets, and show that it can persist after the application of existing calibration methods. To mitigate this issue, we propose strategies for detection, visualization, and quantification of variable-based calibration error. We then examine the limitations of current score-based calibration methods and explore potential modifications. Finally, we discuss the implications of these findings, emphasizing that an understanding of calibration beyond simple aggregate measures is crucial for endeavors such as fairness and model interpretability.

Machine Translation systems can produce different types of errors, some of which are characterized as critical or catastrophic due to the specific negative impact that they can have on users. In this paper we focus on one type of critical error: added toxicity. We evaluate and analyze added toxicity when translating a large evaluation dataset (HOLISTICBIAS, over 472k sentences, covering 13 demographic axes) from English into 164 languages. An automatic toxicity evaluation shows that added toxicity across languages varies from 0% to 5%. The output languages with the most added toxicity tend to be low-resource ones, and the demographic axes with the most added toxicity include sexual orientation, gender and sex, and ability. We also perform human evaluation on a subset of 8 translation directions, confirming the prevalence of true added toxicity. We use a measurement of the amount of source contribution to the translation, where a low source contribution implies hallucination, to interpret what causes toxicity. Making use of the input attributions allows us to explain toxicity, because the source contributions significantly correlate with toxicity for 84% of languages studied. Given our findings, our recommendations to reduce added toxicity are to curate training data to avoid mistranslations, mitigate hallucination and check unstable translations.

Abstract--Continuous Integration (CI) has become a wellestablished software development practice for automatically and continuously integrating code changes during software development. An increasing number of Machine Learning (ML) based approaches for automation of CI phases are being reported in the literature. It is timely and relevant to provide a Systemization of Knowledge (SoK) of ML-based approaches for CI phases. Our systematic analysis also highlights the deficiencies of the existing ML-based solutions that can be improved for advancing the state-of-the-art. Given the variety of employed techniques in applying ML solutions in CI, and growing interest in this domain, it is In recent years, the software development industry has seen necessary to systematically identify state-of-the-art practices a significant shift towards the adoption of Continuous Integration used for automating CI tasks through ML methods.

Reliable application of machine learning-based decision systems in the wild is one of the major challenges currently investigated by the field. A large portion of established approaches aims to detect erroneous predictions by means of assigning confidence scores. This confidence may be obtained by either quantifying the model's predictive uncertainty, learning explicit scoring functions, or assessing whether the input is in line with the training distribution. Curiously, while these approaches all state to address the same eventual goal of detecting failures of a classifier upon real-life application, they currently constitute largely separated research fields with individual evaluation protocols, which either exclude a substantial part of relevant methods or ignore large parts of relevant failure sources. In this work, we systematically reveal current pitfalls caused by these inconsistencies and derive requirements for a holistic and realistic evaluation of failure detection. To demonstrate the relevance of this unified perspective, we present a large-scale empirical study for the first time enabling benchmarking confidence scoring functions w.r.t all relevant methods and failure sources. The revelation of a simple softmax response baseline as the overall best performing method underlines the drastic shortcomings of current evaluation in the abundance of publicized research on confidence scoring. Code and trained models are at https://github.com/IML-DKFZ/fd-shifts.

Random forest is a popular machine learning approach for the analysis of high-dimensional data because it is flexible and provides variable importance measures for the selection of relevant features. However, the complex relationships between the features are usually not considered for the selection and thus also neglected for the characterization of the analysed samples. Here we propose two novel approaches that focus on the mutual impact of features in random forests. Mutual forest impact (MFI) is a relation parameter that evaluates the mutual association of the featurs to the outcome and, hence, goes beyond the analysis of correlation coefficients. Mutual impurity reduction (MIR) is an importance measure that combines this relation parameter with the importance of the individual features. MIR and MFI are implemented together with testing procedures that generate p-values for the selection of related and important features. Applications to various simulated data sets and the comparison to other methods for feature selection and relation analysis show that MFI and MIR are very promising to shed light on the complex relationships between features and outcome. In addition, they are not affected by common biases, e.g. that features with many possible splits or high minor allele frequencies are prefered.

Data sparsity is a main problem hindering the development of code-switching (CS) NLP systems. In this paper, we investigate data augmentation techniques for synthesizing dialectal Arabic-English CS text. We perform lexical replacements using word-aligned parallel corpora where CS points are either randomly chosen or learnt using a sequence-to-sequence model. We compare these approaches against dictionary-based replacements. We assess the quality of the generated sentences through human evaluation and evaluate the effectiveness of data augmentation on machine translation (MT), automatic speech recognition (ASR), and speech translation (ST) tasks. Results show that using a predictive model results in more natural CS sentences compared to the random approach, as reported in human judgements. In the downstream tasks, despite the random approach generating more data, both approaches perform equally (outperforming dictionary-based replacements). Overall, data augmentation achieves 34% improvement in perplexity, 5.2% relative improvement on WER for ASR task, +4.0-5.1 BLEU points on MT task, and +2.1-2.2 BLEU points on ST over a baseline trained on available data without augmentation.

Recently, many improved naive Bayes methods have been developed with enhanced discrimination capabilities. Among them, regularized naive Bayes (RNB) produces excellent performance by balancing the discrimination power and generalization capability. Data discretization is important in naive Bayes. By grouping similar values into one interval, the data distribution could be better estimated. However, existing methods including RNB often discretize the data into too few intervals, which may result in a significant information loss. To address this problem, we propose a semi-supervised adaptive discriminative discretization framework for naive Bayes, which could better estimate the data distribution by utilizing both labeled data and unlabeled data through pseudo-labeling techniques. The proposed method also significantly reduces the information loss during discretization by utilizing an adaptive discriminative discretization scheme, and hence greatly improves the discrimination power of classifiers. The proposed RNB+, i.e., regularized naive Bayes utilizing the proposed discretization framework, is systematically evaluated on a wide range of machine-learning datasets. It significantly and consistently outperforms state-of-the-art NB classifiers.