arXiv subjects

Ronald Tetzlaff

Publications and source records attributed to Ronald Tetzlaff.

11 recordsLinked to original sources

arXiv2026

Identifying Confusion Trends in Concept-based XAI for Multi-Label Classification

Deep Neural Networks (DNNs) deployed in high-risk domains, such as healthcare and autonomous driving, must be not only accurate but also understandable to ensure user trust. In real-world computer vision tasks, these models often operate on complex images containing background noise and are heavily annotated. To make such models explainable, Concept-based Explainable AI (CXAI) methods need to be assessed for their applicability and problem-solving capacity. In this work, we explore CXAI use cases in multi-label classification by training two DNNs, VGG16 and ResNet50, on the 20 most annotated labels in the MS-COCO dataset (Microsoft Common Objects in Context). We apply two CXAI methods, CRP (Concept Relevance Propagation) and CRAFT (Concept Recursive Activation FacTorization), to generate concept-level explanations and investigate the overall evaluations. Our analysis reveals three key findings: (1) CXAI highlights learning weaknesses in DNNs, (2) higher concept distinctiveness reduces label and concept confusion, and (3) environmental concepts expose dataset-induced biases. Our results demonstrate the potential of CXAI to enhance the understanding of model generalizability and to diagnose bias instigated by the dataset.

cs.CV↗

arXiv2024

Block Induced Signature Generative Adversarial Network (BISGAN): Signature Spoofing Using GANs and Their Evaluation

Deep learning is actively being used in biometrics to develop efficient identification and verification systems. Handwritten signatures are a common subset of biometric data for authentication purposes. Generative adversarial networks (GANs) learn from original and forged signatures to generate forged signatures. While most GAN techniques create a strong signature verifier, which is the discriminator, there is a need to focus more on the quality of forgeries generated by the generator model. This work focuses on creating a generator that produces forged samples that achieve a benchmark in spoofing signature verification systems. We use CycleGANs infused with Inception model-like blocks with attention heads as the generator and a variation of the SigCNN model as the base Discriminator. We train our model with a new technique that results in 80% to 100% success in signature spoofing. Additionally, we create a custom evaluation technique to act as a goodness measure of the generated forgeries. Our work advocates generator-focused GAN architectures for spoofing data quality that aid in a better understanding of biometric data generation and evaluation.

cs.CV↗

arXiv2024

Employing Vector Field Techniques on the Analysis of Memristor Cellular Nonlinear Networks Cell Dynamics

This paper introduces an innovative graphical analysis tool for investigating the dynamics of Memristor Cellular Nonlinear Networks (M-CNNs) featuring 2nd-order processing elements, known as M-CNN cells. In the era of specialized hardware catering to the demands of intelligent autonomous systems, the integration of memristors within Cellular Nonlinear Networks (CNNs) has emerged as a promising paradigm due to their exceptional characteristics. However, the standard Dynamic Route Map (DRM) analysis, applicable to 1st-order systems, fails to address the intricacies of 2nd-order M-CNN cell dynamics, as well the 2nd-order DRM (DRM2) exhibits limitations on the graphical illustration of local dynamical properties of the M-CNN cells, e.g. state derivative's magnitude. To address this limitation, we propose a novel integration of M-CNN cell vector field into the cell's phase portrait, enhancing the analysis efficacy and enabling efficient M-CNN cell design. A comprehensive exploration of M-CNN cell dynamics is presented, showcasing the utility of the proposed graphical tool for various scenarios, including bistable and monostable behavior, and demonstrating its superior ability to reveal subtle variations in cell behavior. Through this work, we offer a refined perspective on the analysis and design of M-CNNs, paving the way for advanced applications in edge computing and specialized hardware.

cs.ET↗

arXiv2024

High Frequency Response of Volatile Memristors

In this theoretical study, we focus on the high-frequency response of the electrothermal NbO2-Mott threshold switch, a real-world electronic device, which has been proved to be relevant in several applications and is classified as a volatile memristor. Memristors of this kind, have been shown to exhibit distinctive non-linear behaviors crucial for cutting-edge neuromorphic circuits. In accordance with well-established models for these devices, their resistances depend on their body temperatures, which evolve over time following Newton's Law of Cooling. Here, we demonstrate that HP's NbO2-Mott memristor can manifest up to three distinct steady-state oscillatory behaviors under a suitable high-frequency periodic voltage input, showcasing increased versatility despite its volatile nature. Additionally, when subjected to a high-frequency periodic voltage signal, the device body temperature oscillates with a negligible peak-to-peak amplitude. Since, the temperature remains almost constant over an input cycle, the devices under study behave as linear resistors during each input cycle. Based on these insights, this paper presents analytical equations characterizing the response of the NbO2-Mott memristor to high-frequency voltage inputs, demarcating regions in the state space where distinct initial conditions lead to various asymptotic oscillatory behaviors. Importantly, the mathematical methods introduced in this manuscript are applicable to any volatile electrothermal resistive switch. Additionally, this paper presents analytical equations that accurately reproduce the temperature time-waveform of the studied device during both its transient and steady-state phases when subjected to a zero-mean sinusoidal voltage input oscillating in the high-frequency limit.

cond-mat.mes-hall↗

arXiv2023

A Compact Model of Threshold Switching Devices for Efficient Circuit Simulations

In this paper, we present a new compact model of threshold switching devices which is suitable for efficient circuit-level simulations. First, a macro model, based on a compact transistor based circuit, was implemented in LTSPICE. Then, a descriptive model was extracted and implemented in MATLAB, which is based on the macro model. This macro model was extended to develop a physical model that describes the processes that occur during the threshold switching. The physical model derived comprises a delay structure with few electrical components adjacent to the second junction. The delay model incorporates an internal state variable, which is crucial to transform the descriptive model into a compact model and to parameterize it in terms of electrical parameters that represent the component's behavior. Finally, we applied our model by fitting measured i-v data of an OTS device manufactured by Western Digital Research.

cond-mat.mes-hall↗

arXiv2023

Generalizable Classification of UHF Partial Discharge Signals in Gas-Insulated HVDC Systems Using Neural Networks

Undetected partial discharges (PDs) are a safety critical issue in high voltage (HV) gas insulated systems (GIS). While the diagnosis of PDs under AC voltage is well-established, the analysis of PDs under DC voltage remains an active research field. A key focus of these investigations is the classification of different PD sources to enable subsequent sophisticated analysis. In this paper, we propose and analyze a neural network-based approach for classifying PD signals caused by metallic protrusions and conductive particles on the insulator of HVDC GIS, without relying on pulse sequence analysis features. In contrast to previous approaches, our proposed model can discriminate the studied PD signals obtained at negative and positive potentials, while also generalizing to unseen operating voltage multiples. Additionally, we compare the performance of time- and frequency-domain input signals and explore the impact of different normalization schemes to mitigate the influence of free-space path loss between the sensor and defect location.

cs.LG↗

arXiv2023

Convolutional Neural Networks for Epileptic Seizure Prediction

Epilepsy is the most common neurological disorder and an accurate forecast of seizures would help to overcome the patient's uncertainty and helplessness. In this contribution, we present and discuss a novel methodology for the classification of intracranial electroencephalography (iEEG) for seizure prediction. Contrary to previous approaches, we categorically refrain from an extraction of hand-crafted features and use a convolutional neural network (CNN) topology instead for both the determination of suitable signal characteristics and the binary classification of preictal and interictal segments. Three different models have been evaluated on public datasets with long-term recordings from four dogs and three patients. Overall, our findings demonstrate the general applicability. In this work we discuss the strengths and limitations of our methodology.

cs.LG↗

arXiv2022

Seizure prediction with long-term iEEG recordings: What can we learn from data nonstationarity?

Repeated epileptic seizures impair around 65 million people worldwide and a successful prediction of seizures could significantly help patients suffering from refractory epilepsy. For two dogs with yearlong intracranial electroencephalography (iEEG) recordings, we studied the influence of time series nonstationarity on the performance of seizure prediction using in-house developed machine learning algorithms. We observed a long-term evolution on the scale of weeks or months in iEEG time series that may be represented as switching between certain meta-states. To better predict impending seizures, retraining of prediction algorithms is therefore necessary and the retraining schedule should be adjusted to the change in meta-states. There is evidence that the nature of seizure-free interictal clips also changes with the transition between meta-states, accwhich has been shown relevant for seizure prediction.

q-bio.NC↗

arXiv2021

Coherent False Seizure Prediction in Epilepsy, Coincidence or Providence?

Seizure forecasting using machine learning is possible, but the performance is far from ideal, as indicated by many false predictions and low specificity. Here, we examine false and missing alarms of two algorithms on long-term datasets to show that the limitations are less related to classifiers or features, but rather to intrinsic changes in the data. We evaluated two algorithms on three datasets by computing the correlation of false predictions and estimating the information transfer between both classification methods. For 9 out of 12 individuals both methods showed a performance better than chance. For all individuals we observed a positive correlation in predictions. For individuals with strong correlation in false predictions we were able to boost the performance of one method by excluding test samples based on the results of the second method. Substantially different algorithms exhibit a highly consistent performance and a strong coherency in false and missing alarms. Hence, changing the underlying hypothesis of a preictal state of fixed time length prior to each seizure to a proictal state is more helpful than further optimizing classifiers. The outcome is significant for the evaluation of seizure prediction algorithms on continuous data.

cs.LG↗

arXiv2011

A Memristive Model Compatible with Triplet Rule for Spike-Timing-Dependent-Plasticity

In this paper, we propose an extended version of the memristive STDP model, which is one of the most important and exciting recent discoveries in neuromorphic engineering. The proposed model aims to claim compatibility with another importent STDP rule beyond the pair-based rule, known as the triplet STDP rule. This is an extension of the asynchronous memristive model of Linares-Barranco, et al., capable of explaining the pair-based rule based on the analogy of the synapse to the memristor. The proposed new model is compatible with both the pair-baed and triplet-based rule, by assuming a mechanism of variable thresholds adapting to synaptic potentiation and depression. The dynamical process is governed by ordinary differential equations. The model is an expression of Froemke's principle of suppression for triplet rules and reveals a similar time dependence with that in the suppression STDP model.

physics.bio-ph↗

arXiv2011

Abel Dynamics of Titanium Dioxide Memristor Based on Nonlinear Ionic Drift Model

We give analytical solutions to the titanium dioxide memristor with arbitary order of window functions, which assumes a nonlinear ionic drift model. As the achieved solution, the characteristic curve of state is demonstrated to be a useful tool for determining the operation point, waveform and saturation level. By using this characterizing tool, it is revealed that the same input signal can output completely different u-i orbital dynamics under different initial conditions, which is the uniqueness of memristors. The approach can be regarded as an analogy to using the characteristic curve for the BJT or MOS transisitors. Based on this model, we further propose a class of analytically solvable class of memristive systems that conform to Abel Differential Equations. The equations of state (EOS) of the titanium dioxide memristor based on both linear and nonlinear ionic drift models are typical integrable examples, which can be categorized into this Abel memristor class. This large family of Abel memristive systems offers a frame for obtaining and analyzing the solutions in the closed form, which facilitate their characterization at a more deterministic level.

cond-mat.mes-hall↗