Adversarial examples have proven to threaten speaker identification systems, and several countermeasures against them have been proposed. In this paper, we propose a method to detect the presence of adversarial examples, i.e., a binary classifier distinguishing between benign and adversarial examples. We build upon and extend previous work on attack type classification by exploring new architectures. Additionally, we introduce a method for identifying the victim model on which the adversarial attack is carried out. To achieve this, we generate a new dataset containing multiple attacks performed against various victim models. We achieve an AUC of 0.982 for attack detection, with no more than a 0.03 drop in performance for unknown attacks. Our attack classification accuracy (excluding benign) reaches 86.48% across eight attack types using our LightResNet34 architecture, while our victim model classification accuracy reaches 72.28% across four victim models.

The living-off-the-land (LOTL) offensive methodologies rely on the perpetration of malicious actions through chains of commands executed by legitimate applications, identifiable exclusively by analysis of system logs. LOTL techniques are well hidden inside the stream of events generated by common legitimate activities, moreover threat actors often camouflage activity through obfuscation, making them particularly difficult to detect without incurring in plenty of false alarms, even using machine learning. To improve the performance of models in such an harsh environment, we propose an augmentation framework to enhance and diversify the presence of LOTL malicious activity inside legitimate logs. Guided by threat intelligence, we generate a dataset by injecting attack templates known to be employed in the wild, further enriched by malleable patterns of legitimate activities to replicate the behavior of evasive threat actors. We conduct an extensive ablation study to understand which models better handle our augmented dataset, also manipulated to mimic the presence of model-agnostic evasion and poisoning attacks. Our results suggest that augmentation is needed to maintain high-predictive capabilities, robustness to attack is achieved through specific hardening techniques like adversarial training, and it is possible to deploy near-real-time models with almost-zero false alarms.

Deep extreme classification (XC) aims to train an encoder architecture and an accompanying classifier architecture to tag a data point with the most relevant subset of labels from a very large universe of labels. XC applications in ranking, recommendation and tagging routinely encounter tail labels for which the amount of training data is exceedingly small. Graph convolutional networks (GCN) present a convenient but computationally expensive way to leverage task metadata and enhance model accuracies in these settings. This paper formally establishes that in several use cases, the steep computational cost of GCNs is entirely avoidable by replacing GCNs with non-GCN architectures. The paper notices that in these settings, it is much more effective to use graph data to regularize encoder training than to implement a GCN. Based on these insights, an alternative paradigm RAMEN is presented to utilize graph metadata in XC settings that offers significant performance boosts with zero increase in inference computational costs. RAMEN scales to datasets with up to 1M labels and offers prediction accuracy up to 15% higher on benchmark datasets than state of the art methods, including those that use graph metadata to train GCNs. RAMEN also offers 10% higher accuracy over the best baseline on a proprietary recommendation dataset sourced from click logs of a popular search engine. Code for RAMEN will be released publicly.

Affective states regulate our day to day to function and has a tremendous effect on mental and physical health. Detection of affective states is of utmost importance for mental health monitoring, smart entertainment selection and dynamic workload management. In this paper, we discussed relevant literature on affective state detection using physiology data, the benefits and limitations of different sensors and methods used for collecting physiology data, and our rationale for selecting functional near-infrared spectroscopy. We present the design of an experiment involving nine subjects to evoke the affective states of meditation, amusement and cognitive load and the results of the attempt to classify using machine learning. A mean accuracy of 83.04% was achieved in three class classification with an individual model; 84.39% accuracy was achieved for a group model and 60.57% accuracy was achieved for subject independent model using leave one out cross validation. It was found that prediction accuracy for cognitive load was higher (evoked using a pen and paper task) than the other two classes (evoked using computer bases tasks). To verify that this discrepancy was not due to motor skills involved in the pen and paper task, a second experiment was conducted using four participants and the results of that experiment has also been presented in the paper.

Automated machine learning (AutoML) aims to select and configure machine learning algorithms and combine them into machine learning pipelines tailored to a dataset at hand. For supervised learning tasks, most notably binary and multinomial classification, aka single-label classification (SLC), such AutoML approaches have shown promising results. However, the task of multi-label classification (MLC), where data points are associated with a set of class labels instead of a single class label, has received much less attention so far. In the context of multi-label classification, the data-specific selection and configuration of multi-label classifiers are challenging even for experts in the field, as it is a high-dimensional optimization problem with multi-level hierarchical dependencies. While for SLC, the space of machine learning pipelines is already huge, the size of the MLC search space outnumbers the one of SLC by several orders. In the first part of this thesis, we devise a novel AutoML approach for single-label classification tasks optimizing pipelines of machine learning algorithms, consisting of two algorithms at most. This approach is then extended first to optimize pipelines of unlimited length and eventually configure the complex hierarchical structures of multi-label classification methods. Furthermore, we investigate how well AutoML approaches that form the state of the art for single-label classification tasks scale with the increased problem complexity of AutoML for multi-label classification. In the second part, we explore how methods for SLC and MLC could be configured more flexibly to achieve better generalization performance and how to increase the efficiency of execution-based AutoML systems.

The area under the ROC curve is a common measure that is often used to rank the relative performance of different binary classifiers. However, as has been also previously noted, it can be a measure that ill-captures the benefits of different classifiers when either the true class values or misclassification costs are highly unbalanced between the two classes. We introduce a third dimension to capture these costs, and lift the ROC curve to a ROC surface in a natural way. We study both this surface and introduce the VOROS, the volume over this ROC surface, as a 3D generalization of the 2D area under the ROC curve. For problems where there are only bounds on the expected costs or class imbalances, we restrict consideration to the volume of the appropriate subregion of the ROC surface. We show how the VOROS can better capture the costs of different classifiers on both a classical and a modern example dataset.

Employing Large Language Models (LLM) in various downstream applications such as classification is crucial, especially for smaller companies lacking the expertise and resources required for fine-tuning a model. Fairness in LLMs helps ensure inclusivity, equal representation based on factors such as race, gender and promotes responsible AI deployment. As the use of LLMs has become increasingly prevalent, it is essential to assess whether LLMs can generate fair outcomes when subjected to considerations of fairness. In this study, we introduce a framework outlining fairness regulations aligned with various fairness definitions, with each definition being modulated by varying degrees of abstraction. We explore the configuration for in-context learning and the procedure for selecting in-context demonstrations using RAG, while incorporating fairness rules into the process. Experiments conducted with different LLMs indicate that GPT-4 delivers superior results in terms of both accuracy and fairness compared to other models. This work is one of the early attempts to achieve fairness in prediction tasks by utilizing LLMs through in-context learning.

Human Activity Recognition (HAR) has been extensively studied, with recent emphasis on the implementation of advanced Machine Learning (ML) and Deep Learning (DL) algorithms for accurate classification. This study investigates the efficacy of two ML algorithms, eXtreme Gradient Boosting (XGBoost) and MiniRocket, in the realm of HAR using data collected from smartphone sensors. The experiments are conducted on a dataset obtained from the UCI repository, comprising accelerometer and gyroscope signals captured from 30 volunteers performing various activities while wearing a smartphone. The dataset undergoes preprocessing, including noise filtering and feature extraction, before being utilized for training and testing the classifiers. Monte Carlo cross-validation is employed to evaluate the models' robustness. The findings reveal that both XGBoost and MiniRocket attain accuracy, F1 score, and AUC values as high as 0.99 in activity classification. XGBoost exhibits a slightly superior performance compared to MiniRocket. Notably, both algorithms surpass the performance of other ML and DL algorithms reported in the literature for HAR tasks. Additionally, the study compares the computational efficiency of the two algorithms, revealing XGBoost's advantage in terms of training time. Furthermore, the performance of MiniRocket, which achieves accuracy and F1 values of 0.94, and an AUC value of 0.96 using raw data and utilizing only one channel from the sensors, highlights the potential of directly leveraging unprocessed signals. It also suggests potential advantages that could be gained by utilizing sensor fusion or channel fusion techniques. Overall, this research sheds light on the effectiveness and computational characteristics of XGBoost and MiniRocket in HAR tasks, providing insights for future studies in activity recognition using smartphone sensor data.

We propose AFTNet, a novel network-constraint survival analysis method based on the Weibull accelerated failure time (AFT) model solved by a penalized likelihood approach for variable selection and estimation. When using the log-linear representation, the inference problem becomes a structured sparse regression problem for which we explicitly incorporate the correlation patterns among predictors using a double penalty that promotes both sparsity and grouping effect. Moreover, we establish the theoretical consistency for the AFTNet estimator and present an efficient iterative computational algorithm based on the proximal gradient descent method. Finally, we evaluate AFTNet performance both on synthetic and real data examples.

Identification and appropriate handling of inconsistencies in data at deployment time is crucial to reliably use machine learning models. While recent data-centric methods are able to identify such inconsistencies with respect to the training set, they suffer from two key limitations: (1) suboptimality in settings where features exhibit statistical independencies, due to their usage of compressive representations and (2) lack of localization to pin-point why a sample might be flagged as inconsistent, which is important to guide future data collection. We solve these two fundamental limitations using directed acyclic graphs (DAGs) to encode the training set's features probability distribution and independencies as a structure. Our method, called DAGnosis, leverages these structural interactions to bring valuable and insightful data-centric conclusions. DAGnosis unlocks the localization of the causes of inconsistencies on a DAG, an aspect overlooked by previous approaches. Moreover, we show empirically that leveraging these interactions (1) leads to more accurate conclusions in detecting inconsistencies, as well as (2) provides more detailed insights into why some samples are flagged.