High-accurate machine learning (ML) image classifiers cannot guarantee that they will not fail at operation. Thus, their deployment in safety-critical applications such as autonomous vehicles is still an open issue. The use of fault tolerance mechanisms such as safety monitors is a promising direction to keep the system in a safe state despite errors of the ML classifier. As the prediction from the ML is the core information directly impacting safety, many works are focusing on monitoring the ML model itself. Checking the efficiency of such monitors in the context of safety-critical applications is thus a significant challenge. Therefore, this paper aims at establishing a baseline framework for benchmarking monitors for ML image classifiers. Furthermore, we propose a framework covering the entire pipeline, from data generation to evaluation. Our approach measures monitor performance with a broader set of metrics than usually proposed in the literature. Moreover, we benchmark three different monitor approaches in 79 benchmark datasets containing five categories of out-of-distribution data for image classifiers: class novelty, noise, anomalies, distributional shifts, and adversarial attacks. Our results indicate that these monitors are no more accurate than a random monitor. We also release the code of all experiments for reproducibility.

This PhD thesis contains several contributions to the field of statistical causal modeling. Statistical causal models are statistical models embedded with causal assumptions that allow for the inference and reasoning about the behavior of stochastic systems affected by external manipulation (interventions). This thesis contributes to the research areas concerning the estimation of causal effects, causal structure learning, and distributionally robust (out-of-distribution generalizing) prediction methods. We present novel and consistent linear and non-linear causal effects estimators in instrumental variable settings that employ data-dependent mean squared prediction error regularization. Our proposed estimators show, in certain settings, mean squared error improvements compared to both canonical and state-of-the-art estimators. We show that recent research on distributionally robust prediction methods has connections to well-studied estimators from econometrics. This connection leads us to prove that general K-class estimators possess distributional robustness properties. We, furthermore, propose a general framework for distributional robustness with respect to intervention-induced distributions. In this framework, we derive sufficient conditions for the identifiability of distributionally robust prediction methods and present impossibility results that show the necessity of several of these conditions. We present a new structure learning method applicable in additive noise models with directed trees as causal graphs. We prove consistency in a vanishing identifiability setup and provide a method for testing substructure hypotheses with asymptotic family-wise error control that remains valid post-selection. Finally, we present heuristic ideas for learning summary graphs of nonlinear time-series models.

Fast developing artificial intelligence (AI) technology has enabled various applied systems deployed in the real world, impacting people's everyday lives. However, many current AI systems were found vulnerable to imperceptible attacks, biased against underrepresented groups, lacking in user privacy protection, etc., which not only degrades user experience but erodes the society's trust in all AI systems. In this review, we strive to provide AI practitioners a comprehensive guide towards building trustworthy AI systems. We first introduce the theoretical framework of important aspects of AI trustworthiness, including robustness, generalization, explainability, transparency, reproducibility, fairness, privacy preservation, alignment with human values, and accountability. We then survey leading approaches in these aspects in the industry. To unify the current fragmented approaches towards trustworthy AI, we propose a systematic approach that considers the entire lifecycle of AI systems, ranging from data acquisition to model development, to development and deployment, finally to continuous monitoring and governance. In this framework, we offer concrete action items to practitioners and societal stakeholders (e.g., researchers and regulators) to improve AI trustworthiness. Finally, we identify key opportunities and challenges in the future development of trustworthy AI systems, where we identify the need for paradigm shift towards comprehensive trustworthy AI systems.

It is meant in a way that we generate multiple new training data for statistical forecasting methods like ARIMA or triple exponential smoothing (Holt-Winters method etc.) to improve forecasting accuracy. It is called bootstrapping, and after applying the forecasting method on each new time series, forecasts are then aggregated by average or median – then it is bagging – bootstrap aggregating. It is proofed by multiple methods, e.g. in regression, that bagging helps improve predictive accuracy – in methods like classical bagging, random forests, gradient boosting methods and so on. The bagging methods for time series forecasting were used also in the latest M4 forecasting competition. For residential electricity consumption (load) time series (as used in my previous blog posts), I proposed three new bootstrapping methods for time series forecasting methods.

Background: Single-cell RNA sequencing (scRNA-seq) yields valuable insights about gene expression and gives critical information about complex tissue cellular composition. In the analysis of single-cell RNA sequencing, the annotations of cell subtypes are often done manually, which is time-consuming and irreproducible. Garnett is a cell-type annotation software based the on elastic net method. Besides cell-type annotation, supervised machine learning methods can also be applied to predict other cell phenotypes from genomic data. Despite the popularity of such applications, there is no existing study to systematically investigate the performance of those supervised algorithms in various sizes of scRNA-seq data sets. Methods and Results: This study evaluates 13 popular supervised machine learning algorithms to classify cell phenotypes, using published real and simulated data sets with diverse cell sizes. The benchmark contained two parts. In the first part, we used real data sets to assess the popular supervised algorithms' computing speed and cell phenotype classification performance. The classification performances were evaluated using AUC statistics, F1-score, precision, recall, and false-positive rate. In the second part, we evaluated gene selection performance using published simulated data sets with a known list of real genes. Conclusion: The study outcomes showed that ElasticNet with interactions performed best in small and medium data sets. NB was another appropriate method for medium data sets. In large data sets, XGB works excellent. Ensemble algorithms were not significantly superior to individual machine learning methods. Adding interactions to ElasticNet can help, and the improvement was significant in small data sets.

Line of sight satellite systems, unmanned aerial vehicles, high-altitude platforms, and microwave links that operate on frequency bands such as Ka-band or higher are extremely susceptible to rain. Thus, rain fade forecasting for these systems is critical because it allows the system to switch between ground gateways proactively before a rain fade event to maintain seamless service. Although empirical, statistical, and fade slope models can predict rain fade to some extent, they typically require statistical measurements of rain characteristics in a given area and cannot be generalized to a large scale system. Furthermore, such models typically predict near-future rain fade events but are incapable of forecasting far into the future, making proactive resource management more difficult. In this paper, a deep learning (DL)-based architecture is proposed that forecasts future rain fade using satellite and radar imagery data as well as link power measurements. Furthermore, the data preprocessing and architectural design have been thoroughly explained and multiple experiments have been conducted. Experiments show that the proposed DL architecture outperforms current state-of-the-art machine learning-based algorithms in rain fade forecasting in the near and long term. Moreover, the results indicate that radar data with weather condition information is more effective for short-term prediction, while satellite data with cloud movement information is more effective for long-term predictions.

Empirical observations often consist of anomalies (or outliers) that contaminate the data. Accurate identification of anomalous samples is crucial for the success of downstream data analysis tasks. To automatically identify anomalies, we propose a new type of autoencoder (AE) which we term Probabilistic Robust autoencoder (PRAE). PRAE is designed to simultaneously remove outliers and identify a low-dimensional representation for the inlier samples. We first describe Robust AE (RAE) as a model that aims to split the data to inlier samples from which a low dimensional representation is learned via an AE, and anomalous (outlier) samples that are excluded as they do not fit the low dimensional representation. Robust AE minimizes the reconstruction of the AE while attempting to incorporate as many observations as possible. This could be realized by subtracting from the reconstruction term an $\ell_0$ norm counting the number of selected observations. Since the $\ell_0$ norm is not differentiable, we propose two probabilistic relaxations for the RAE approach and demonstrate that they can effectively identify anomalies. We prove that the solution to PRAE is equivalent to the solution of RAE and demonstrate using extensive simulations that PRAE is at par with state-of-the-art methods for anomaly detection.

While the 10-year survival rate for localized prostate cancer patients is very good (>98%), side effects of treatment may limit quality of life significantly. Erectile dysfunction (ED) is a common burden associated with increasing age as well as prostate cancer treatment. Although many studies have investigated the factors affecting erectile dysfunction (ED) after prostate cancer treatment, only limited studies have investigated whether ED can be predicted before the start of treatment. The advent of machine learning (ML) based prediction tools in oncology offers a promising approach to improve accuracy of prediction and quality of care. Predicting ED may help aid shared decision making by making the advantages and disadvantages of certain treatments clear, so that a tailored treatment for an individual patient can be chosen. This study aimed to predict ED at 1-year and 2-year post-diagnosis based on patient demographics, clinical data and patient-reported outcomes (PROMs) measured at diagnosis.

The tremendous success of generative models in recent years raises the question whether they can also be used to perform classification. Generative models have been used as adversarially robust classifiers on simple datasets such as MNIST, but this robustness has not been observed on more complex datasets like CIFAR-10. Additionally, on natural image datasets, previous results have suggested a trade-off between the likelihood of the data and classification accuracy. In this work, we investigate score-based generative models as classifiers for natural images. We show that these models not only obtain competitive likelihood values but simultaneously achieve state-of-the-art classification accuracy for generative classifiers on CIFAR-10. Nevertheless, we find that these models are only slightly, if at all, more robust than discriminative baseline models on out-of-distribution tasks based on common image corruptions. Similarly and contrary to prior results, we find that score-based are prone to worst-case distribution shifts in the form of adversarial perturbations. Our work highlights that score-based generative models are closing the gap in classification accuracy compared to standard discriminative models. While they do not yet deliver on the promise of adversarial and out-of-domain robustness, they provide a different approach to classification that warrants further research.

An increasingly common use case for machine learning models is augmenting the abilities of human decision makers. For classification tasks where neither the human or model are perfectly accurate, a key step in obtaining high performance is combining their individual predictions in a manner that leverages their relative strengths. In this work, we develop a set of algorithms that combine the probabilistic output of a model with the class-level output of a human. We show theoretically that the accuracy of our combination model is driven not only by the individual human and model accuracies, but also by the model's confidence. Empirical results on image classification with CIFAR-10 and a subset of ImageNet demonstrate that such human-model combinations consistently have higher accuracies than the model or human alone, and that the parameters of the combination method can be estimated effectively with as few as ten labeled datapoints.