Weaknesses: 1. Figure 2 is very confusing to me. Figure (2a) seems to train an individual classifier for each domain. Figure (2b) seems also to train an individual classifier for each domain but also require the agreement across all classifiers for all samples. My question is what is the difference between Figure 2b and the method only train one classifier for all domains? It seems to me that training several classifiers and alignment them with a loss is same as only training one classifier. Furthermore, I agree that alignment classifiers can align features across domains as well, but the previous methods also use some distance loss or adversarial learning to align classifiers or features, which will reach a similar performance.

As LLMs increase in accessibility, LLM-generated texts have proliferated across several fields, such as scientific, academic, and creative writing. However, LLMs are not created equally; they may have different architectures and training datasets. Thus, some LLMs may be more challenging to detect than others. Using two datasets spanning four total writing domains, we train AI-generated (AIG) text classifiers using the LibAUC library - a deep learning library for training classifiers with imbalanced datasets. Our results in the Deepfake Text dataset show that AIG-text detection varies across domains, with scientific writing being relatively challenging. In the Rewritten Ivy Panda (RIP) dataset focusing on student essays, we find that the OpenAI family of LLMs was substantially difficult for our classifiers to distinguish from human texts. Additionally, we explore possible factors that could explain the difficulties in detecting OpenAI-generated texts.

Safe overtakes in trucks are crucial to prevent accidents, reduce congestion, and ensure efficient traffic flow, making early prediction essential for timely and informed driving decisions. Accordingly, we investigate the detection of truck overtakes from CAN data. Three classifiers, Artificial Neural Networks (ANN), Random Forest, and Support Vector Machines (SVM), are employed for the task. Our analysis covers up to 10 seconds before the overtaking event, using an overlapping sliding window of 1 second to extract CAN features. We observe that the prediction scores of the overtake class tend to increase as we approach the overtake trigger, while the no-overtake class remain stable or oscillates depending on the classifier. Thus, the best accuracy is achieved when approaching the trigger, making early overtaking prediction challenging. The classifiers show good accuracy in classifying overtakes (Recall/TPR > 93%), but accuracy is suboptimal in classifying no-overtakes (TNR typically 80-90% and below 60% for one SVM variant). We further combine two classifiers (Random Forest and linear SVM) by averaging their output scores. The fusion is observed to improve no-overtake classification (TNR > 92%) at the expense of reducing overtake accuracy (TPR). However, the latter is kept above 91% near the overtake trigger. Therefore, the fusion balances TPR and TNR, providing more consistent performance than individual classifiers.

Multimodal integration is a key component of allowing robots to perceive the world. Multimodality comes with multiple challenges that have to be considered, such as how to integrate and fuse the data. In this paper, we compare different possibilities of fusing visual, tactile and proprioceptive data. The data is directly recorded on the NICOL robot in an experimental setup in which the robot has to classify containers and their content. Due to the different nature of the containers, the use of the modalities can wildly differ between the classes. We demonstrate the superiority of multimodal solutions in this use case and evaluate three fusion strategies that integrate the data at different time steps. We find that the accuracy of the best fusion strategy is 15% higher than the best strategy using only one singular sense.

Improving and guaranteeing the robustness of deep learning models has been a topic of intense research. Ensembling, which combines several classifiers to provide a better model, has shown to be beneficial for generalisation, uncertainty estimation, calibration, and mitigating the effects of concept drift. However, the impact of ensembling on certified robustness is less well understood. In this work, we generalise Lipschitz continuity by introducing S-Lipschitz classifiers, which we use to analyse the theoretical robustness of ensembles. Our results are precise conditions when ensembles of robust classifiers are more robust than any constituent classifier, as well as conditions when they are less robust.

Computational pathology tasks have some unique characterises such as multi-gigapixel images, tedious and frequently uncertain annotations, and unavailability of large number of cases [13]. To address some of these issues, we present Deep Fastfood Ensembles - a simple, fast and yet effective method for combining deep features pooled from popular CNN models pre-trained on totally different source domains (e.g., natural image objects) and projected onto diverse dimensions using random projections, the so-called Fastfood [11]. The final ensemble output is obtained by a consensus of simple individual classifiers, each of which is trained on a different collection of random basis vectors. This offers extremely fast and yet effective solution, especially when training times and domain labels are of the essence. We demonstrate the effectiveness of the proposed deep fastfood ensemble learning as compared to the state-of-the-art methods for three different tasks in histopathology image analysis.

Signal-background classification is a central problem in High-Energy Physics (HEP), that plays a major role for the discovery of new fundamental particles. A recent method -- the Parametric Neural Network (pNN) -- leverages multiple signal mass hypotheses as an additional input feature to effectively replace a whole set of individual classifiers, each providing (in principle) the best response for the corresponding mass hypothesis. In this work we aim at deepening the understanding of pNNs in light of real-world usage. We discovered several peculiarities of parametric networks, providing intuition, metrics, and guidelines to them. We further propose an alternative parametrization scheme, resulting in a new parametrized neural network architecture: the AffinePNN; along with many other generally applicable improvements, like the balanced training procedure. Finally, we extensively and empirically evaluate our models on the HEPMASS dataset, along its imbalanced version (called HEPMASS-IMB) we provide here for the first time, to further validate our approach. Provided results are in terms of the impact of the proposed design decisions, classification performance, and interpolation capability, as well.

Digital twins have recently gained significant interest in simulation, optimization, and predictive maintenance of Industrial Control Systems (ICS). Recent studies discuss the possibility of using digital twins for intrusion detection in industrial systems. Accordingly, this study contributes to a digital twin-based security framework for industrial control systems, extending its capabilities for simulation of attacks and defense mechanisms. Four types of process-aware attack scenarios are implemented on a standalone open-source digital twin of an industrial filling plant: command injection, network Denial of Service (DoS), calculated measurement modification, and naive measurement modification. A stacked ensemble classifier is proposed as the real-time intrusion detection, based on the offline evaluation of eight supervised machine learning algorithms. The designed stacked model outperforms previous methods in terms of F1-Score and accuracy, by combining the predictions of various algorithms, while it can detect and classify intrusions in near real-time (0.1 seconds). This study also discusses the practicality and benefits of the proposed digital twin-based security framework.

We propose an information-fusion approach based on belief functions to combine convolutional neural networks. In this approach, several pre-trained DS-based CNN architectures extract features from input images and convert them into mass functions on different frames of discernment. A fusion module then aggregates these mass functions using Dempster's rule. An end-to-end learning procedure allows us to fine-tune the overall architecture using a learning set with soft labels, which further improves the classification performance. The effectiveness of this approach is demonstrated experimentally using three benchmark databases.

Consider an ensemble of $k$ individual classifiers whose accuracies are known. Upon receiving a test point, each of the classifiers outputs a predicted label and a confidence in its prediction for this particular test point. In this paper, we address the question of whether we can determine the accuracy of the ensemble. Surprisingly, even when classifiers are combined in the statistically optimal way in this setting, the accuracy of the resulting ensemble classifier cannot be computed from the accuracies of the individual classifiers-as would be the case in the standard setting of confidence weighted majority voting. We prove tight upper and lower bounds on the ensemble accuracy. We explicitly construct the individual classifiers that attain the upper and lower bounds: specialists and generalists. Our theoretical results have very practical consequences: (1) If we use ensemble methods and have the choice to construct our individual (independent) classifiers from scratch, then we should aim for specialist classifiers rather than generalists. (2) Our bounds can be used to determine how many classifiers are at least required to achieve a desired ensemble accuracy. Finally, we improve our bounds by considering the mutual information between the true label and the individual classifier's output.