In human-AI collaboration systems for critical applications, in order to ensure minimal error, users should set an operating point based on model confidence to determine when the decision should be delegated to human experts. Samples for which model confidence is lower than the operating point would be manually analysed by experts to avoid mistakes. Such systems can become truly useful only if they consider two aspects: models should be confident only for samples for which they are accurate, and the number of samples delegated to experts should be minimized. The latter aspect is especially crucial for applications where available expert time is limited and expensive, such as healthcare. The trade-off between the model accuracy and the number of samples delegated to experts can be represented by a curve that is similar to an ROC curve, which we refer to as confidence operating characteristic (COC) curve. In this paper, we argue that deep neural networks should be trained by taking into account both accuracy and expert load and, to that end, propose a new complementary loss function for classification that maximizes the area under this COC curve. This promotes simultaneously the increase in network accuracy and the reduction in number of samples delegated to humans. We perform experiments on multiple computer vision and medical image datasets for classification. Our results demonstrate that the proposed loss improves classification accuracy and delegates less number of decisions to experts, achieves better out-of-distribution samples detection and on par calibration performance compared to existing loss functions.

The detection of state-sponsored trolls operating in influence campaigns on social media is a critical and unsolved challenge for the research community, which has significant implications beyond the online realm. To address this challenge, we propose a new AI-based solution that identifies troll accounts solely through behavioral cues associated with their sequences of sharing activity, encompassing both their actions and the feedback they receive from others. Our approach does not incorporate any textual content shared and consists of two steps: First, we leverage an LSTM-based classifier to determine whether account sequences belong to a state-sponsored troll or an organic, legitimate user. Second, we employ the classified sequences to calculate a metric named the "Troll Score", quantifying the degree to which an account exhibits troll-like behavior. To assess the effectiveness of our method, we examine its performance in the context of the 2016 Russian interference campaign during the U.S. Presidential election. Our experiments yield compelling results, demonstrating that our approach can identify account sequences with an AUC close to 99% and accurately differentiate between Russian trolls and organic users with an AUC of 91%. Notably, our behavioral-based approach holds a significant advantage in the ever-evolving landscape, where textual and linguistic properties can be easily mimicked by Large Language Models (LLMs): In contrast to existing language-based techniques, it relies on more challenging-to-replicate behavioral cues, ensuring greater resilience in identifying influence campaigns, especially given the potential increase in the usage of LLMs for generating inauthentic content. Finally, we assessed the generalizability of our solution to various entities driving different information operations and found promising results that will guide future research.

This paper provides a comprehensive survey of bias mitigation methods for achieving fairness in Machine Learning (ML) models. We collect a total of 341 publications concerning bias mitigation for ML classifiers. These methods can be distinguished based on their intervention procedure (i.e., pre-processing, in-processing, post-processing) and the technique they apply. We investigate how existing bias mitigation methods are evaluated in the literature. In particular, we consider datasets, metrics and benchmarking. Based on the gathered insights (e.g., What is the most popular fairness metric? How many datasets are used for evaluating bias mitigation methods?), we hope to support practitioners in making informed choices when developing and evaluating new bias mitigation methods.

Matrix-variate normal mixture models are a powerful statistical tool used to represent complex data structures that involve matrices, such as multivariate time series, spatial data, and image data. Detecting outliers in matrix-variate normal mixture models is crucial for identifying anomalous observations that deviate significantly from the underlying distribution. Outliers can provide valuable insights into data quality issues, anomalies, or unexpected patterns. Outliers, and their treatment, is a long-studied topic in the field of applied statistics. The problem of handling outliers in multivariate clustering has been studied in several contexts including work by García-Escudero et al. (2008), Punzo and McNicholas (2016), Punzo et al. (2020), and Clark and McNicholas (2023).

Wide heterogeneity exists in cancer patients' survival, ranging from a few months to several decades. To accurately predict clinical outcomes, it is vital to build an accurate predictive model that relates patients' molecular profiles with patients' survival. With complex relationships between survival and high-dimensional molecular predictors, it is challenging to conduct non-parametric modeling and irrelevant predictors removing simultaneously. In this paper, we build a kernel Cox proportional hazards semi-parametric model and propose a novel regularized garrotized kernel machine (RegGKM) method to fit the model. We use the kernel machine method to describe the complex relationship between survival and predictors, while automatically removing irrelevant parametric and non-parametric predictors through a LASSO penalty. An efficient high-dimensional algorithm is developed for the proposed method. Comparison with other competing methods in simulation shows that the proposed method always has better predictive accuracy. We apply this method to analyze a multiple myeloma dataset and predict patients' death burden based on their gene expressions. Our results can help classify patients into groups with different death risks, facilitating treatment for better clinical outcomes.

Data class imbalance is a common problem in classification problems, where minority class samples are often more important and more costly to misclassify in a classification task. Therefore, it is very important to solve the data class imbalance classification problem. The SMART dataset exhibits an evident class imbalance, comprising a substantial quantity of healthy samples and a comparatively limited number of defective samples. This dataset serves as a reliable indicator of the disc's health status. In this paper, we obtain the best balanced disk SMART dataset for a specific classification model by mixing and integrating the data synthesised by multivariate generative adversarial networks (GAN) to balance the disk SMART dataset at the data level; and combine it with genetic algorithms to obtain higher disk fault classification prediction accuracy on a specific classification model.

Machine learning (ML) - based software systems are rapidly gaining adoption across various domains, making it increasingly essential to ensure they perform as intended. This report presents best practices for the Test and Evaluation (T&E) of ML-enabled software systems across its lifecycle. We categorize the lifecycle of ML-enabled software systems into three stages: component, integration and deployment, and post-deployment. At the component level, the primary objective is to test and evaluate the ML model as a standalone component. Next, in the integration and deployment stage, the goal is to evaluate an integrated ML-enabled system consisting of both ML and non-ML components. Finally, once the ML-enabled software system is deployed and operationalized, the T&E objective is to ensure the system performs as intended. Maintenance activities for ML-enabled software systems span the lifecycle and involve maintaining various assets of ML-enabled software systems. Given its unique characteristics, the T&E of ML-enabled software systems is challenging. While significant research has been reported on T&E at the component level, limited work is reported on T&E in the remaining two stages. Furthermore, in many cases, there is a lack of systematic T&E strategies throughout the ML-enabled system's lifecycle. This leads practitioners to resort to ad-hoc T&E practices, which can undermine user confidence in the reliability of ML-enabled software systems. New systematic testing approaches, adequacy measurements, and metrics are required to address the T&E challenges across all stages of the ML-enabled system lifecycle.

Automatic crash bucketing is a crucial phase in the software development process for efficiently triaging bug reports. It generally consists in grouping similar reports through clustering techniques. However, with real-time streaming bug collection, systems are needed to quickly answer the question: What are the most similar bugs to a new one?, that is, efficiently find near-duplicates. It is thus natural to consider nearest neighbors search to tackle this problem and especially the well-known locality-sensitive hashing (LSH) to deal with large datasets due to its sublinear performance and theoretical guarantees on the similarity search accuracy. Surprisingly, LSH has not been considered in the crash bucketing literature. It is indeed not trivial to derive hash functions that satisfy the so-called locality-sensitive property for the most advanced crash bucketing metrics. Consequently, we study in this paper how to leverage LSH for this task. To be able to consider the most relevant metrics used in the literature, we introduce DeepLSH, a Siamese DNN architecture with an original loss function, that perfectly approximates the locality-sensitivity property even for Jaccard and Cosine metrics for which exact LSH solutions exist. We support this claim with a series of experiments on an original dataset, which we make available.

Constant evolution and the emergence of new cyberattacks require the development of advanced techniques for defense. This paper aims to measure the impact of a supervised filter (classifier) in network anomaly detection. We perform our experiments by employing a hybrid anomaly detection approach in network flow data. For this purpose, we extended a state-of-the-art autoencoder-based anomaly detection method by prepending a binary classifier acting as a prefilter for the anomaly detector. The method was evaluated on the publicly available real-world dataset UGR'16. Our empirical results indicate that the hybrid approach does offer a higher detection rate of known attacks than a standalone anomaly detector while still retaining the ability to detect zero-day attacks. Employing a supervised binary prefilter has increased the AUC metric by over 11%, detecting 30% more attacks while keeping the number of false positives approximately the same.

Recent years have witnessed a surge in research on machine learning for combinatorial optimization since learning-based approaches can outperform traditional heuristics and approximate exact solvers at a lower computation cost. However, most existing work on supervised neural combinatorial optimization focuses on TSP instances with a fixed number of cities and requires large amounts of training samples to achieve a good performance, making them less practical to be applied to realistic optimization scenarios. This work aims to develop a data-driven graph representation learning method for solving travelling salesman problems (TSPs) with various numbers of cities. To this end, we propose an edge-aware graph autoencoder (EdgeGAE) model that can learn to solve TSPs after being trained on solution data of various sizes with an imbalanced distribution. We formulate the TSP as a link prediction task on sparse connected graphs. A residual gated encoder is trained to learn latent edge embeddings, followed by an edge-centered decoder to output link predictions in an end-to-end manner. To improve the model's generalization capability of solving large-scale problems, we introduce an active sampling strategy into the training process. In addition, we generate a benchmark dataset containing 50,000 TSP instances with a size from 50 to 500 cities, following an extremely scale-imbalanced distribution, making it ideal for investigating the model's performance for practical applications. We conduct experiments using different amounts of training data with various scales, and the experimental results demonstrate that the proposed data-driven approach achieves a highly competitive performance among state-of-the-art learning-based methods for solving TSPs.