Oscar Deniz

This work presents a two-model framework for synthesizing human-like mouse trajectories under user-specified spatial and temporal targets. The first model, DeepLog, is a Deep Learning approach based on the Sigma-Lognormal model that decomposes expert mouse trajectories into constituent strokes. These strokes encode the information required to reconstruct the original trajectories and provide a compact representation for subsequent generation. DeepLog provides a robust and efficient alternative to the iDeLog reference algorithm, statistically outperforming it in both reconstruction accuracy and robustness. Specifically, the signal-to-noise ratios for shape and velocity improve by factors of 1.43 and 1.33, respectively, while the Hausdorff distance, discrete Fréchet distance, and Dynamic Time Warping improve by factors of 1.13, 1.22, and 2.03, respectively. The second model uses the stroke representations obtained with DeepLog as expert demonstrations within an Inverse Reinforcement Learning (IRL) framework. This generator learns a reward function that guides the synthesis of human-like mouse movements and enables the generation of trajectories conditioned on both spatial destinations and temporal constraints. Overall, the proposed framework combines stroke-based trajectory representation with adversarial imitation learning to generate flexible, human-like mouse movements beyond the specific spatial and temporal configurations observed in the demonstrations.

Deep learning (DL) has significantly transformed computer vision, demonstrating remarkable achievements and extensive real‐world applications. However, recent studies have highlighted a critical vulnerability of DL models to adversarial examples (AE), where slight perturbations in input data can lead to erroneous outputs. We observe that the behaviour of the AE is similar to a chaotic system, where a minor change in the input leads to a significantly different output. In response, we propose a novel approach for detecting and categorizing adversarial inputs encountered by classification neural networks. The proposed approach focuses on extracting statistical profiles, termed as chaotic feature vectors (CFVs), from a collection of features derived from ordinal patterns (OP). In this work, the proposed AE detection method is tested on seven attack methods and three image datasets including MNIST, FMNIST and CIFAR10. The results indicate that CFVs exhibit promising capabilities in discerning AE against various types of adversarial attacks on different datasets. This advancement lays the foundation for devising attack mitigation strategies, thereby enhancing the robustness and security of DL models in the face of adversarial threats.

In the realm of deep learning, deep neural networks (DNNs) have recently propelled significant advancements in image classification applications. However, these DNN models are vulnerable to adversarial examples (AE), which are crafted by introducing imperceptible perturbations to legitimate samples, leading the models to make incorrect predictions. Time-series (TS) data are prevalent in real-world applications and naturally occurring phenomena. To recover and infer missing information, TS analysis has been extensively explored as a foundational research problem. The introduction of imperceptible perturbations induces irregularities in TS, resulting in chaotic behavior. Chaos, marked by a lack of organization and structure, is expected to play diverse functional roles in dynamical systems when analyzed through TS. By applying chaos theory to TS analysis, we aim to identify chaotic behavior to detect the subtle input perturbations added by the adversarial attacks. Chaos theory faces significant challenges when applied to the analysis or computation of chaotic systems due to the minimal and subtle nature of the perturbations introduced by adversarial attacks. Nonetheless, detecting chaos amidst adversarial perturbations is crucial for defending against such attacks. In this paper, we utilize features derived from ordinal patterns (OP) through TS analysis to detect the presence or absence of chaos, even under sophisticated attack methods. As an initial application of our pipeline, we exploit the edge-preserving capabilities of the guided filter (GF) and the spatial proximity properties of the Hilbert space-filling curve (HSFC) to represent the TS of input images. Experimental results show that our proposed method effectively identifies adversarial perturbations across various attacks on diverse datasets, including MNIST, Fashion-MNIST (FMNIST), CIFAR-10 and CIFAR-100.