Refine
Document Type
- Conference Proceeding (194)
- Article (reviewed) (98)
- Article (unreviewed) (21)
- Letter to Editor (16)
- Doctoral Thesis (8)
- Book (7)
- Patent (7)
- Part of a Book (3)
- Moving Images (1)
- Other (1)
Conference Type
- Konferenzartikel (169)
- Konferenz-Abstract (19)
- Sonstiges (5)
- Konferenz-Poster (1)
Language
- English (357) (remove)
Is part of the Bibliography
- yes (357) (remove)
Keywords
- RoboCup (12)
- Deep Leaning (11)
- Machine Learning (11)
- Götz von Berlichingen (7)
- neuroprosthetics (7)
- Blockchain (6)
- machine learning (6)
- Heart rhythm model (5)
- Modeling and simulation (5)
- neural networks (5)
Institute
- Fakultät Elektrotechnik, Medizintechnik und Informatik (EMI) (ab 04/2019) (357) (remove)
Open Access
- Open Access (180)
- Closed Access (110)
- Closed (62)
- Bronze (31)
- Gold (28)
- Diamond (22)
- Grün (3)
- Hybrid (2)
Biodegradable metals have entered the implant market in recent years, but still do not show fully satisfactory degradation behaviour and mechanical properties. In contrast, it has been shown that pure molybdenum has an excellent combination of the required properties in this respect. We report on PM based screen printing of thin-walled molybdenum tubes as a processing step for medical stent manufacture. We also present data on the in vivo degradation and biocompatibility of molybdenum. The degradation of molybdenum wires implanted in the aorta of rats was evaluated by SEM and EDX. Biocompatibility was assessed by histological investigation of organs and analysis of molybdenum levels in tissue extracts and body fluids. Degradation rates of up to 13.5 μm/y were observed after 12 months. No histological changes or elevated molybdenum levels in organ tissues were observed. In summary, the results further underline that molybdenum is a highly promising biodegradable metallic material.
Titanium and stainless steel are commonly known as osteosynthesis materials with high strength and good biocompatibility. However, they have the big disadvantage that a second operation for hardware removal is necessary. Although resorbable systems made of polymers or magnesium are increasingly used, they show some severe adverse foreign body reactions or unsatisfying degradation behavior. Therefore, we started to investigate molybdenum as a potential new biodegradable material for osteosynthesis in craniomaxillofacial surgery. To characterize molybdenum as a biocompatible material, we performed in vitro assays in accordance with ISO Norm 10993-5. In four different experimental setups, we showed that pure molybdenum and molybdenum rhenium alloys do not lead to cytotoxicity in human and mouse fibroblasts. We also examined the degradation behavior of molybdenum by carrying out long-term immersion tests (up to 6 months) with molybdenum sheet metal. We showed that molybdenum has sufficient mechanical stability over at least 6 months for implants on the one hand and is subject to very uniform degradation on the other. The results of our experiments are very promising for the development of new resorbable osteosynthesis materials for craniomaxillofacial surgery based on molybdenum.
Digital transformation strengthens the interconnection of companies in order to develop optimized and better customized, cross-company business models. These models require secure, reliable, and traceable evidence and monitoring of contractually agreed information to gain trust between stakeholders. Blockchain technology using smart contracts allows the industry to establish trust and automate cross-company business processes without the risk of losing data control. A typical cross-company industry use case is equipment maintenance. Machine manufacturers and service providers offer maintenance for their machines and tools in order to achieve high availability at low costs. The aim of this chapter is to demonstrate how maintenance use cases are attempted by utilizing hyperledger fabric for building a chain of trust by hardened evidence logging of the maintenance process to achieve legal certainty. Contracts are digitized into smart contracts automating business that increase the security and mitigate the error-proneness of the business processes.
Blockchain-IIoT integration into industrial processes promises greater security, transparency, and traceability. However, this advancement faces significant storage and scalability issues with existing blockchain technologies. Each peer in the blockchain network maintains a full copy of the ledger which is updated through consensus. This full replication approach places a burden on the storage space of the peers and would quickly outstrip the storage capacity of resource-constrained IIoT devices. Various solutions utilizing compression, summarization or different storage schemes have been proposed in literature. The use of cloud resources for blockchain storage has been extensively studied in recent years. Nonetheless, block selection remains a substantial challenge associated with cloud resources and blockchain integration. This paper proposes a deep reinforcement learning (DRL) approach as an alternative to solving the block selection problem, which involves identifying the blocks to be transferred to the cloud. We propose a DRL approach to solve our problem by converting the multi-objective optimization of block selection into a Markov decision process (MDP). We design a simulated blockchain environment for training and testing our proposed DRL approach. We utilize two DRL algorithms, Advantage Actor-Critic (A2C), and Proximal Policy Optimization (PPO) to solve the block selection problem and analyze their performance gains. PPO and A2C achieve 47.8% and 42.9% storage reduction on the blockchain peer compared to the full replication approach of conventional blockchain systems. The slowest DRL algorithm, A2C, achieves a run-time 7.2 times shorter than the benchmark evolutionary algorithms used in earlier works, which validates the gains introduced by the DRL algorithms. The simulation results further show that our DRL algorithms provide an adaptive and dynamic solution to the time-sensitive blockchain-IIoT environment.
The desire to connect more and more devices and to make them more intelligent and more reliable, is driving the needs for the Internet of Things more than ever. Such IoT edge systems require sound security measures against cyber-attacks, since they are interconnected, spatially distributed, and operational for an extended period of time. One of the most important requirements for the security in many industrial IoT applications is the authentication of the devices. In this paper, we present a mutual authentication protocol based on Physical Unclonable Functions, where challenge-response pairs are used for both device and server authentication. Moreover, a session key can be derived by the protocol in order to secure the communication channel. We show that our protocol is secure against machine learning, replay, man-in-the-middle, cloning, and physical attacks. Moreover, it is shown that the protocol benefits from a smaller computational, communication, storage, and hardware overhead, compared to similar works.
In recent years, Physical Unclonable Functions (PUFs) have gained significant attraction in the Internet of Things (IoT) for security applications such as cryptographic key generation and entity authentication. PUFs extract the uncontrollable production characteristics of physical devices to generate unique fingerprints for security applications. One common approach for designing PUFs is exploiting the intrinsic features of sensors and actuators such as MEMS elements, which typically exist in IoT devices. This work presents the Cantilever-PUF, a PUF based on a specific MEMS device – Aluminum Nitride (AlN) piezoelectric cantilever. We show the variations of electrical parameters of AlN cantilevers such as resonance frequency, electrical conductivity, and quality factor, as a result of uncontrollable manufacturing process variations. These variations, along with high thermal and chemical stability, and compatibility with silicon technology, makes AlN cantilever a decent candidate for PUF design. We present a cantilever design, which magnifies the effect of manufacturing process variations on electrical parameters. In order to verify our findings, the simulation results of the Monte Carlo method are provided. The results verify the eligibility of AlN cantilever to be used as a basic PUF device for security applications. We present an architecture, in which the designed Cantilever-PUF is used as a security anchor for PUF-enabled device authentication as well as communication encryption.
Spatially Distributed Wireless Networks (SDWN) are one of the basic technologies for the Internet of Things (IoT) and (Industrial) Internet of Things (IIoT) applications. These SDWN for many of these applications has strict requirements such as low cost, simple installation and operations, and high potential flexibility and mobility. Among the different Narrowband Wireless Wide Area Networking (NBWWAN) technologies, which are introduced to address these categories of wireless networking requirements, Narrowband Internet of Things (NB-IoT) is getting more traction due to attractive system parameters, energy-saving mode of operation with low data rates and bandwidth, and its applicability in 5G use cases. Since several technologies are available and because the underlying use cases come with various requirements, it is essential to perform a systematic comparative analysis of competing technologies to choose the right technology. It is also important to perform testing during different phases of the system development life cycle. This paper describes the systematic test environment for automated testing of radio communication and systematic measurements of the performance of NB-IoT.
In this paper, we study the runtime performance of symmetric cryptographic algorithms on an embedded ARM Cortex-M4 platform. Symmetric cryptographic algorithms can serve to protect the integrity and optionally, if supported by the algorithm, the confidentiality of data. A broad range of well-established algorithms exists, where the different algorithms typically have different properties and come with different computational complexity. On deeply embedded systems, the overhead imposed by cryptographic operations may be significant. We execute the algorithms AES-GCM, ChaCha20-Poly1305, HMAC-SHA256, KMAC, and SipHash on an STM32 embedded microcontroller and benchmark the execution times of the algorithms as a function of the input lengths.
In recent years, the topic of embedded machine learning has become very popular in AI research. With the help of various compression techniques such as pruning, quantization and others compression techniques, it became possible to run neural networks on embedded devices. These techniques have opened up a whole new application area for machine learning. They range from smart products such as voice assistants to smart sensors that are needed in robotics. Despite the achievements in embedded machine learning, efficient algorithms for training neural networks in constrained domains are still lacking. Training on embedded devices will open up further fields of applications. Efficient training algorithms would enable federated learning on embedded devices, in which the data remains where it was collected, or retraining of neural networks in different domains. In this paper, we summarize techniques that make training on embedded devices possible. We first describe the need and requirements for such algorithms. Then we examine existing techniques that address training in resource-constrained environments as well as techniques that are also suitable for training on embedded devices, such as incremental learning. At the end, we also discuss which problems and open questions still need to be solved in these areas.
The EREMI project is a 2-year project funded under the ERASMUS+ framework programme and its team has developed and will validate an advanced higher education program, including life-long learning, on the interdisciplinary topic of resource efficiency in manufacturing industries and the overall system optimization of low or not digitized physical infrastructure. All of these will be achieved by applying IoT technologies towards efficient industrial systems, and by utilizing a high-level educated human capital on these economically, politically, and technically crucial and highly relevant topics for the rapidly developing industries and economies of intensively economically and industrially transforming countries - Bulgaria, North Macedonia, and Romania. Efficiency will be attained by utilizing the experience and expertise of the involved German partner organisation.
The integration of Internet of Things devices onto the Blockchain implies an increase in the transactions that occur on the Blockchain, thus increasing the storage requirements.
A solution approach is to leverage cloud resources for storing blocks within the chain. The paper, therefore, proposes two solutions to this problem. The first being an improved hybrid architecture design which uses containerization to create a side chain on a fog node for the devices connected to it and an Advanced Time‑variant Multi‑objective Particle Swarm Optimization Algorithm (AT‑MOPSO) for determining the optimal number of blocks that should be transferred to the cloud for storage. This algorithm uses time‑variant weights for the velocity of the particle swarm optimization and the non‑dominated sorting and mutation schemes from NSGA‑III. The proposed algorithm was compared with results from the original MOPSO algorithm, the Strength Pareto Evolutionary Algorithm (SPEA‑II), and the Pareto Envelope‑based Selection Algorithm with region‑based selection (PESA‑II), and NSGA‑III. The proposed AT‑MOPSO showed better results than the aforementioned MOPSO algorithms in cloud storage cost and query probability optimization. Importantly, AT‑MOPSO achieved 52% energy efficiency compared to NSGA‑III.
To show how this algorithm can be applied to a real‑world Blockchain system, the BISS industrial Blockchain architecture was adapted and modified to show how the AT‑MOPSO can be used with existing Blockchain systems and the benefits it provides.
Recent work has investigated the distributions of learned convolution filters through a large-scale study containing hundreds of heterogeneous image models. Surprisingly, on average, the distributions only show minor drifts in comparisons of various studied dimensions including the learned task, image domain, or dataset. However, among the studied image domains, medical imaging models appeared to show significant outliers through "spikey" distributions, and, therefore, learn clusters of highly specific filters different from other domains. Following this observation, we study the collected medical imaging models in more detail. We show that instead of fundamental differences, the outliers are due to specific processing in some architectures. Quite the contrary, for standardized architectures, we find that models trained on medical data do not significantly differ in their filter distributions from similar architectures trained on data from other domains. Our conclusions reinforce previous hypotheses stating that pre-training of imaging models can be done with any kind of diverse image data.
Currently, many theoretical as well as practically relevant questions towards the transferability and robustness of Convolutional Neural Networks (CNNs) remain unsolved. While ongoing research efforts are engaging these problems from various angles, in most computer vision related cases these approaches can be generalized to investigations of the effects of distribution shifts in image data. In this context, we propose to study the shifts in the learned weights of trained CNN models. Here we focus on the properties of the distributions of dominantly used 3×3 convolution filter kernels. We collected and publicly provide a dataset with over 1.4 billion filters from hundreds of trained CNNs, using a wide range of datasets, architectures, and vision tasks. In a first use case of the proposed dataset, we can show highly relevant properties of many publicly available pre-trained models for practical applications: I) We analyze distribution shifts (or the lack thereof) between trained filters along different axes of meta-parameters, like visual category of the dataset, task, architecture, or layer depth. Based on these results, we conclude that model pre-training can succeed on arbitrary datasets if they meet size and variance conditions. II) We show that many pre-trained models contain degenerated filters which make them less robust and less suitable for fine-tuning on target applications. Data & Project website: https://github.com/paulgavrikov/cnn-filter-db.
Deep learning models are intrinsically sensitive to distribution shifts in the input data. In particular, small, barely perceivable perturbations to the input data can force models to make wrong predictions with high confidence. An common defense mechanism is regularization through adversarial training which injects worst-case perturbations back into training to strengthen the decision boundaries, and to reduce overfitting. In this context, we perform an investigation of 3 × 3 convolution filters that form in adversarially- trained models. Filters are extracted from 71 public models of the ℓ ∞ -RobustBench CIFAR-10/100 and ImageNet1k leaderboard and compared to filters extracted from models built on the same architectures but trained without robust regularization. We observe that adversarially-robust models appear to form more diverse, less sparse, and more orthogonal convolution filters than their normal counterparts. The largest differences between robust and normal models are found in the deepest layers, and the very first convolution layer, which consistently and predominantly forms filters that can partially eliminate perturbations, irrespective of the architecture.
The conversion of space heating for private households to climate-neutral energy sources is an essential component of the energy transition, as this sector as of 2018 was responsible for 9.4 % of Germany’s carbon dioxide emissions. In addition to reducing demand through better insulation, the use of heat pumps fed with electricity from renewable energy sources, such as on-site photovoltaics (PV) systems, is an important solution approach.
Advanced energy management and control can help to make optimal use of such heating systems. Optimal here can e.g. refer to maximizing self-consumption of self-generated PV power, extended component lifetime or a grid-friendly behavior that avoids load peaks. A powerful method for this is model predictive control (MPC), which calculates optimal schedules for the controllable influence variables based on models of the system dynamics, current measurements of system states and predictions of future external influence parameters.
In this paper, we will discuss three different use cases that show how artificial intelligence can contribute to the realization of such an MPC-based energy management and control system. This will be done using the example of a real inhabited single family home that has provided the necessary data for this purpose and where the methods are implemented and tested. The heating system consists of an air-water heat pump with direct condensation, a thermal stratified storage tank, a pellet burner and a heating rod and provides both heating and hot water. The house generates a significant portion of its electricity needs through a rooftop PV system.
In automotive parking scenario, where the curb shall be detected and classified to be traversable or not, radars play an important role. There are different approaches already proposed in other works to estimate the target height. This paper assesses and compares two methods. The first is based on Angle of Arrival (AoA) estimation of input signals of multiple antennas using the Multiple-Input-Multiple-Output (MIMO) principle. The second method uses the geometry in multipath propagation of the radar echo signal for one antenna input. In this work a modified method of calculation of the curb height based on the second method is proposed. The theory of approach is mathematically proved and effectiveness is demonstrated by evaluation of measurements with a 77 GHz Frequency Modulated Continuous Wave (FMCW) radar. In order to evaluate the performance of the introduced method the mean square error (MSE) is used in the proposed scenario. This method, using only one antenna input, produced up to 3.4 times better results for curb height detection in comparison with former methods.
In this study, various imaging algorithms for the localization of objects have been investigated. Therefore, an Ultra-Wideband (UWB) radar based experimental setup with a circular antenna array is designed as part of this work. This concept could be particularly useful in microwave medical imaging applications. In order to validate its applicability in microwave imaging, different imaging algorithms have been evaluated and compared by means of our experimental setup. Accurate imaging results have been achieved with our system under multiple test-scenarios.
In this study, an approach to a microwave-based radar system for the localization of objects has been proposed. This could be particularly useful in microwave imaging applications such as cardiac catheter detection. An experimental system is defined and realized with the selection of an appropriate antenna design. Hardware control functions and different imaging algorithms are implemented as well. The functionality of this measurement setup has been analyzed considering multiple testscenarios and it is proved to be capable of locating multiple objects as well as expanded objects.
In this paper, we propose a unified approach for network pruning and one-shot neural architecture search (NAS) via group sparsity. We first show that group sparsity via the recent Proximal Stochastic Gradient Descent (ProxSGD) algorithm achieves new state-of-the-art results for filter pruning. Then, we extend this approach to operation pruning, directly yielding a gradient-based NAS method based on group sparsity. Compared to existing gradient-based algorithms such as DARTS, the advantages of this new group sparsity approach are threefold. Firstly, instead of a costly bilevel optimization problem, we formulate the NAS problem as a single-level optimization problem, which can be optimally and efficiently solved using ProxSGD with convergence guarantees. Secondly, due to the operation-level sparsity, discretizing the network architecture by pruning less important operations can be safely done without any performance degradation. Thirdly, the proposed approach finds architectures that are both stable and well-performing on a variety of search spaces and datasets.
Despite the success of convolutional neural networks (CNNs) in many academic benchmarks for computer vision tasks, their application in the real-world is still facing fundamental challenges. One of these open problems is the inherent lack of robustness, unveiled by the striking effectiveness of adversarial attacks. Adversarial training (AT) is often considered as a remedy to train more robust networks. In this paper, we empirically analyze a variety of adversarially trained models that achieve high robust accuracies when facing state-of-the-art attacks and we show that AT has an interesting side-effect: it leads to models that are significantly less overconfident with their decisions even on clean data than non-robust models. Further, our analysis of robust models shows that not only AT but also the model's building blocks (like activation functions and pooling) have a strong influence on the models' prediction confidences.
Harnessing the overall benefits of the latest advancements in artificial intelligence (AI) requires the extensive collaboration of academia and industry. These collaborations promote innovation and growth while enforcing the practical usefulness of newer technologies in real life. The purpose of this article is to outline the challenges faced during cross-collaboration between academia and industry. These challenges are also inspected with the help of an ongoing project titled “Quality Assurance of Machine Learning Applications” (Q-AMeLiA), in which three universities cooperate with five industry partners to make the product risk of AI-based products visible. Further, we discuss the hurdles and the key challenges in machine learning (ML) technology transformation from academia to industry based on robustness, simplicity, and safety. These challenges are an outcome of the lack of common standards, metrics, and missing regulatory considerations when state-of-the-art (SOTA) technology is developed in academia. The use of biased datasets involves ethical concerns that might lead to unfair outcomes when the ML model is deployed in production. The advancement of AI in small and medium sized enterprises (SMEs) requires more in terms of common tandardization of concepts rather than algorithm breakthroughs. In this paper, in addition to the general challenges, we also discuss domain specific barriers for five different domains i.e., object detection, hardware benchmarking, continual learning, action recognition, and industrial process automation, and highlight the steps necessary for successfully managing the cross-sectoral collaborations between academia and industry.
In this work, we explore three deep learning algorithms apply to seismic interpolation: deep prior image (DPI), standard, and generative adversarial networks (GAN). The standard and GAN approaches rely on a dataset of complete and decimated seismic images for the training process, while the DPI method learns from a decimated image itself, without training images. We carry out two main experiments, considering 10%, 30%, and 50% of regular and irregular decimation. The first tests the optimal situation for the GAN and the standard approaches, where training and testing images are from the same dataset. The second tests the ability of GAN and standard methods to learn simultaneously from three datasets, and generalize to a fourth dataset not used during training. The standard method provides the best results in the first experiment, when the training distribution is similar to the testing one. In this situation, the DPI approach reports the second best results. In the second experiment, the standard method shows the ability to learn simultaneously and effectively three data distributions for the regular case. In the irregular case, the DPI approach is more effective. The GAN approach is the less effective of the three deep learning methods in both experiments.
Seismic data has often missing traces due to technical acquisition or economical constraints. A compete dataset is crucial in several processing and inversion techniques. Deep learning algorithms, based on convolutional neural networks (CNNs), have shown alternative solutions that overcome limitation of traditional interpolation methods e.g. data regularity, linearity assumption, etc. There are two different paradigms of CNN methods for seismic interpolation. The first one, so-called deep prior interpolation (DPI), trains a CNN to map random noise to a complete seismic image using only the decimated image itself. The second one, referred as standard deep learning method, trains a CNN to map a decimated seismic image into a complete one using a dataset of complete and artificially decimated images. Within this research, we systematically compare the performance of both methods for different quantities of regular and irregular missing traces using 4 datasets. We evaluate the results of both methods using 5 well-known metrics. We found that DPI method performs better than the standard method if the percentage of missing traces is low (10%) and otherwise if the level of decimation is high (50%).
Recently, RobustBench (Croce et al. 2020) has become a widely recognized benchmark for the adversarial robustness of image
classification networks. In it’s most commonly reported sub-task, RobustBench evaluates and ranks the adversarial robustness of trained neural networks on CIFAR10 under AutoAttack (Croce and Hein 2020b) with l∞ perturbations limited to ϵ = 8/255. With leading scores of the currently best performing models of around 60% of the baseline, it is fair to characterize this benchmark to be quite challenging. Despite it’s general acceptance in recent literature, we aim to foster discussion about the suitability of RobustBench as a key indicator for robustness which could be generalized to practical applications. Our line of argumentation against this is two-fold and supported by excessive experiments presented in this paper: We argue that I) the alternation of data by AutoAttack with l∞, ϵ = 8/255 is unrealistically strong, resulting in close to perfect detection rates of adversarial samples even by simple detection algorithms and human observers.
We also show that other attack methods are much harder to detect while achieving similar success rates. II) That results on low resolution data sets like CIFAR10 do not generalize well to higher resolution images as gradient based attacks appear to become even more detectable with increasing resolutions.
Many commonly well-performing convolutional neural network models have shown to be susceptible to input data perturbations, indicating a low model robustness. Adversarial attacks are thereby specifically optimized to reveal model weaknesses, by generating small, barely perceivable image perturbations that flip the model prediction. Robustness against attacks can be gained for example by using adversarial examples during training, which effectively reduces the measurable model attackability. In contrast, research on analyzing the source of a model’s vulnerability is scarce. In this paper, we analyze adversarially trained, robust models in the context of a specifically suspicious network operation, the downsampling layer, and provide evidence that robust models have learned to downsample more accurately and suffer significantly less from aliasing than baseline models.
Estimating the Robustness of Classification Models by the Structure of the Learned Feature-Space
(2022)
Over the last decade, the development of deep image classification networks has mostly been driven by the search for the best performance in terms of classification accuracy on standardized benchmarks like ImageNet. More recently, this focus has been expanded by the notion of model robustness, \ie the generalization abilities of models towards previously unseen changes in the data distribution. While new benchmarks, like ImageNet-C, have been introduced to measure robustness properties, we argue that fixed testsets are only able to capture a small portion of possible data variations and are thus limited and prone to generate new overfitted solutions. To overcome these drawbacks, we suggest to estimate the robustness of a model directly from the structure of its learned feature-space. We introduce robustness indicators which are obtained via unsupervised clustering of latent representations from a trained classifier and show very high correlations to the model performance on corrupted test data.
Aerosol particles play an important role in the climate system by absorbing and scattering radiation and influencing cloud properties. They are also one of the biggest sources of uncertainty for climate modeling. Many climate models do not include aerosols in sufficient detail due to computational constraints. To represent key processes, aerosol microphysical properties and processes have to be accounted for. This is done in the ECHAM-HAM (European Center for Medium-Range Weather Forecast-Hamburg-Hamburg) global climate aerosol model using the M7 microphysics, but high computational costs make it very expensive to run with finer resolution or for a longer time. We aim to use machine learning to emulate the microphysics model at sufficient accuracy and reduce the computational cost by being fast at inference time. The original M7 model is used to generate data of input–output pairs to train a neural network (NN) on it. We are able to learn the variables’ tendencies achieving an average R² score of 77.1%. We further explore methods to inform and constrain the NN with physical knowledge to reduce mass violation and enforce mass positivity. On a Graphics processing unit (GPU), we achieve a speed-up of up to over 64 times faster when compared to the original model.
Many commonly well-performing convolutional neural network models have shown to be susceptible to input data perturbations, indicating a low model robustness. To reveal model weaknesses, adversarial attacks are specifically optimized to generate small, barely perceivable image perturbations that flip the model prediction. Robustness against attacks can be gained by using adversarial examples during training, which in most cases reduces the measurable model attackability. Unfortunately, this technique can lead to robust overfitting, which results in non-robust models. In this paper, we analyze adversarially trained, robust models in the context of a specific network operation, the downsampling layer, and provide evidence that robust models have learned to downsample more accurately and suffer significantly less from downsampling artifacts, aka. aliasing, than baseline models. In the case of robust overfitting, we observe a strong increase in aliasing and propose a novel early stopping approach based on the measurement of aliasing.
Despite the success of convolutional neural networks (CNNs) in many academic benchmarks for computer vision tasks, their application in the real-world is still facing fundamental challenges. One of these open problems is the inherent lack of robustness, unveiled by the striking effectiveness of adversarial attacks. Current attack methods are able to manipulate the network's prediction by adding specific but small amounts of noise to the input. In turn, adversarial training (AT) aims to achieve robustness against such attacks and ideally a better model generalization ability by including adversarial samples in the trainingset. However, an in-depth analysis of the resulting robust models beyond adversarial robustness is still pending. In this paper, we empirically analyze a variety of adversarially trained models that achieve high robust accuracies when facing state-of-the-art attacks and we show that AT has an interesting side-effect: it leads to models that are significantly less overconfident with their decisions, even on clean data than non-robust models. Further, our analysis of robust models shows that not only AT but also the model's building blocks (like activation functions and pooling) have a strong influence on the models' prediction confidences. Data & Project website: https://github.com/GeJulia/robustness_confidences_evaluation
Over the last years, Convolutional Neural Networks (CNNs) have been the dominating neural architecture in a wide range of computer vision tasks. From an image and signal processing point of view, this success might be a bit surprising as the inherent spatial pyramid design of most CNNs is apparently violating basic signal processing laws, i.e. Sampling Theorem in their down-sampling operations. However, since poor sampling appeared not to affect model accuracy, this issue has been broadly neglected until model robustness started to receive more attention. Recent work in the context of adversarial attacks and distribution shifts, showed after all, that there is a strong correlation between the vulnerability of CNNs and aliasing artifacts induced by poor down-sampling operations. This paper builds on these findings and introduces an aliasing free down-sampling operation which can easily be plugged into any CNN architecture: FrequencyLowCut pooling. Our experiments show, that in combination with simple and Fast Gradient Sign Method (FGSM) adversarial training, our hyper-parameter free operator substantially improves model robustness and avoids catastrophic overfitting. Our code is available at https://github.com/GeJulia/flc_pooling
Featherweight Go (FG) is a minimal core calculus that includes essential Go features such as overloaded methods and interface types. The most straightforward semantic description of the dynamic behavior of FG programs is to resolve method calls based on run-time type information. A more efficient approach is to apply a type-directed translation scheme where interface-values are replaced by dictionaries that contain concrete method definitions. Thus, method calls can be resolved by a simple lookup of the method definition in the dictionary. Establishing that the target program obtained via the type-directed translation scheme preserves the semantics of the original FG program is an important task.
To establish this property we employ logical relations that are indexed by types to relate source and target programs. We provide rigorous proofs and give a detailed discussion of the many subtle corners that we have encountered including the need for a step index due to recursive inter- faces and method definitions.
Solar energy plays a central role in the energy transition. Clouds generate locally large fluctuations in the generation output of photovoltaic systems, which is a major problem for energy systems such as microgrids, among others. For an optimal design of a power system, this work analyzed the variability using a spatially distributed sensor network at Stuttgart Airport. It has been shown that the spatial distribution partially reduces the variability of solar radiation. A tool was also developed to estimate the output power of photovoltaic systems using irradiation time series and assumptions about the photovoltaic sites. For days with high fluctuations of the estimated photovoltaic power, different energy system scenarios were investigated. It was found the approach can be used to have a more realistic representation of aggregated PV power taking spatial smoothing into account and that the resulting PV power generation profiles provide a good basis for energy system design considerations like battery sizing.
When people with hearing loss are provided with different devices in each ear, these devices usually have different processing latencies. This leads to static temporal offsets between both ears in the order of several milliseconds. This thesis measured effects of such offsets in stimulation timing on mechanisms of binaural hearing, such as sound localization and speech understanding in noise in hearing-impaired and normal-hearing listeners.
In this paper, a concept for an anthropomorphic replacement hand cast with silicone with an integrated sensory feedback system is presented. In order to construct the personalized replacement hand, a 3D scan of a healthy hand was used to create a 3D-printed mold using computer-aided design (CAD). To allow for movement of the index and middle fingers, a motorized orthosis was used. Information about the applied force for grasping and the degree of flexion of the fingers is registered using two pressure sensors and one bending sensor in each movable finger. To integrate the sensors and additional cavities for increased flexibility, the fingers were cast in three parts, separately from the rest of the hand. A silicone adhesive (Silpuran 4200) was examined to combine the individual parts afterwards. For this, tests with different geometries were carried out. Furthermore, different test series for the secure integration of the sensors were performed, including measurements of the registered information of the sensors. Based on these findings, skin-toned individual fingers and a replacement hand with integrated sensors were created. Using Silpuran 4200, it was possible to integrate the needed cavities and to place the sensors securely into the hand while retaining full flexion using a motorized orthosis. The measurements during different loadings and while grasping various objects proved that it is possible to realize such a sensory feedback system in a replacement hand. As a result, it can be stated that the cost-effective realization of a personalized, anthropomorphic replacement hand with an integrated sensory feedback system is possible using 3D scanning and 3D printing. By integrating smaller sensors, the risk of damaging the sensors through movement could be decreased.
We consider the local group of agents for exchanging the time-series data value and computing the approximation of the mean value of all agents. An agent represented by a node knows all local neighbor nodes in the same group. The node has the contact information of other nodes in other groups. The nodes interact with each other in synchronous rounds to exchange the updated time-series data value using the random call communication model. The amount of data exchanged between agent-based sensors in the local group network affects the accuracy of the aggregation function results. At each time step, the agent-based sensor can update the input data value and send the updated data value to the group head node. The group head node sends the updated data value to all group members in the same group. Grouping nodes in peer-to-peer networks show an improvement in Mean Squared Error (MSE).
Positioning mobile systems with high accuracy is a prerequisite for intelligent autonomous behavior, both in industrial environments and in field robotics. This paper describes the setup of a robotic platform and its use for the evaluation of simultaneous localization and mapping (SLAM) algorithms. A configuration using a mobile robot Husky A200, and a LiDAR (light detection and ranging) sensor was used to implement the setup. For verification of the proposed setup, different scan matching methods for odometry determination in indoor and outdoor environments are tested. An assessment of the accuracy of the baseline 3D-SLAM system and the selected evaluation system is presented by comparing different scenarios and test situations. It was shown that the hdl_graph_slam in combination with the LiDAR OS1 and the scan matching algorithms FAST_GICP and FAST_VGICP achieves good mapping results with accuracies up to 2 cm.
Electrode modelling and simulation of diagnostic and pulmonary vein isolation in atrial fibrillation
(2022)
An Overview of Technologies for Improving Storage Efficiency in Blockchain-Based IIoT Applications
(2022)
Since the inception of blockchain-based cryptocurrencies, researchers have been fascinated with the idea of integrating blockchain technology into other fields, such as health and manufacturing. Despite the benefits of blockchain, which include immutability, transparency, and traceability, certain issues that limit its integration with IIoT still linger. One of these prominent problems is the storage inefficiency of the blockchain. Due to the append-only nature of the blockchain, the growth of the blockchain ledger inevitably leads to high storage requirements for blockchain peers. This poses a challenge for its integration with the IIoT, where high volumes of data are generated at a relatively faster rate than in applications such as financial systems. Therefore, there is a need for blockchain architectures that deal effectively with the rapid growth of the blockchain ledger. This paper discusses the problem of storage inefficiency in existing blockchain systems, how this affects their scalability, and the challenges that this poses to their integration with IIoT. This paper explores existing solutions for improving the storage efficiency of blockchain–IIoT systems, classifying these proposed solutions according to their approaches and providing insight into their effectiveness through a detailed comparative analysis and examination of their long-term sustainability. Potential directions for future research on the enhancement of storage efficiency in blockchain–IIoT systems are also discussed.
Note: In lieu of an abstract, this is an excerpt from the first page.
Recently, we reported the three-dimensional computer-aided design (3D-CAD) reconstruction of the first “Iron Hand” of the famous Franconian knight, Götz von Berlichingen (1480–1562), who lost his right hand by a cannon ball splinter injury in 1504 in the War of the Succession of Landshut [...]
In asymmetric treatment of hearing loss, processing latencies of the modalities typically differ. This often alters the reference interaural time difference (ITD) (i.e., the ITD at 0° azimuth) by several milliseconds. Such changes in reference ITD have shown to influence sound source localization in bimodal listeners provided with a hearing aid (HA) in one and a cochlear implant (CI) in the contralateral ear. In this study, the effect of changes in reference ITD on speech understanding, especially spatial release from masking (SRM) in normal-hearing subjects was explored. Speech reception thresholds (SRT) were measured in ten normal-hearing subjects for reference ITDs of 0, 1.75, 3.5, 5.25 and 7 ms with spatially collocated (S0N0) and spatially separated (S0N90) sound sources. Further, the cues for separation of target and masker were manipulated to measure the effect of a reference ITD on unmasking by A) ITDs and interaural level differences (ILDs), B) ITDs only and C) ILDs only. A blind equalization-cancellation (EC) model was applied to simulate all measured conditions. SRM decreased significantly in conditions A) and B) when the reference ITD was increased: In condition A) from 8.8 dB SNR on average at 0 ms reference ITD to 4.6 dB at 7 ms, in condition B) from 5.5 dB to 1.1 dB. In condition C) no significant effect was found. These results were accurately predicted by the applied EC-model. The outcomes show that interaural processing latency differences should be considered in asymmetric treatment of hearing loss.
The increase of the Internet of Things (IoT) calls for secure solutions for industrial applications. The security of IoT can be potentially improved by blockchain. However, blockchain technology suffers scalability issues which hinders integration with IoT. Solutions to blockchain’s scalability issues, such as minimizing the computational complexity of consensus algorithms or blockchain storage requirements, have received attention. However, to realize the full potential of blockchain in IoT, the inefficiencies of its inter-peer communication must also be addressed. For example, blockchain uses a flooding technique to share blocks, resulting in duplicates and inefficient bandwidth usage. Moreover, blockchain peers use a random neighbor selection (RNS) technique to decide on other peers with whom to exchange blockchain data. As a result, the peer-to-peer (P2P) topology formation limits the effective achievable throughput. This paper provides a survey on the state-of-the-art network structures and communication mechanisms used in blockchain and establishes the need for network-based optimization. Additionally, it discusses the blockchain architecture and its layers categorizes existing literature into the layers and provides a survey on the state-of-the-art optimization frameworks, analyzing their effectiveness and ability to scale. Finally, this paper presents recommendations for future work.
Commercial simulators can only reproduce electrocardiograms (ECG) of the normal and diseased heart rhythm in a simplified waveform and with a low number of channels. With the presented project, the variety of digitally archived ECGs, recorded during electrophysiological examinations, should be made usable as original analogue signals for research and teaching purposes by the development of a special printed circuit board for the mini-computer “Raspberry-Pi “.
Industrial companies can use blockchain to assist them in resolving their trust and security issues. In this research, we provide a fully distributed blockchain-based architecture for industrial IoT, relying on trust management and reputation to enhance nodes’ trustworthiness. The purpose of this contribution is to introduce our system architecture to show how to secure network access for users with dynamic authorization management. All decisions in the system are made by trustful nodes’ consensus and are fully distributed. The remarkable feature of this system architecture is that the influence of the nodes’ power is lowered depending on their Proof of Work (PoW) and Proof of Stake (PoS), and the nodes’ significance and authority is determined by their behavior in the network.
This impact is based on game theory and an incentive mechanism for reputation between nodes. This system design can be used on legacy machines, which means that security and distributed systems
can be put in place at a low cost on industrial systems. While there are no numerical results yet, this work, based on the open questions regarding the majority problem and the proposed solutions based on a game-theoretic mechanism and a trust management system, points to what and how industrial IoT and existing blockchain frameworks that are focusing only on the power of PoW and PoS can be secured more effectively.
The present invention relates to open-loop and closed-loop control units for extracorporeal circulatory support, to systems comprising such an open-loop and closed-loop control unit, and to corresponding methods. An open-loop and closed-loop control unit (10) for extracorporeal circulatory support is proposed, which is configured to receive a measurement of an ECG signal (12) of a supported patient over a predefined period of time, wherein the ECG signal (12) comprises multiple data points for each time point within a heart cycle. The open-loop and closed-loop control unit (10) comprises an evaluation unit (100) which is configured to evaluate the data points for at least one time point in a spatial and/or temporal manner and to determine at least one amplitude change (14) within the heart cycle based on the evaluated data points. The open-loop and closed-loop control unit (10) is further configured to output an open-loop and/or closed-loop signal (16) for extracorporeal circulatory support at a predefined point in time after the at least one amplitude change (14).
The present invention relates to open-loop and closed-loop control units for extracorporeal circulatory support, to systems comprising such an open-loop and closed-loop control unit, and to corresponding methods. An open-loop and closed-loop control unit (10) for extracorporeal circulatory support is proposed, which is configured to receive a measurement of an ECG signal (12) of a supported patient over a predefined period of time, wherein the ECG signal (12) comprises multiple data points for each time point within a heart cycle. The open-loop and closed-loop control unit (10) comprises an evaluation unit (100) which is configured to evaluate the data points for at least one time point in a spatial and/or temporal manner and to determine at least one amplitude change (14) within the heart cycle based on the evaluated data points. The open-loop and closed-loop control unit (10) is further configured to output an open-loop and/or closed-loop signal (16) for extracorporeal circulatory support at a predefined point in time after the at least one amplitude change (14).