Refine
Document Type
- Conference Proceeding (211)
- Article (reviewed) (127)
- Article (unreviewed) (27)
- Patent (23)
- Letter to Editor (16)
- Book (12)
- Contribution to a Periodical (12)
- Doctoral Thesis (12)
- Report (11)
- Part of a Book (10)
- Moving Images (1)
- Other (1)
- Working Paper (1)
Conference Type
- Konferenzartikel (183)
- Konferenz-Abstract (21)
- Sonstiges (5)
- Konferenz-Poster (2)
Language
- English (357)
- German (104)
- Other language (1)
- Multiple languages (1)
- Russian (1)
Is part of the Bibliography
- yes (464) (remove)
Keywords
- Machine Learning (13)
- RoboCup (12)
- Deep Leaning (11)
- Götz von Berlichingen (9)
- neuroprosthetics (7)
- Blockchain (6)
- machine learning (6)
- Heart rhythm model (5)
- Herzrhythmusmodell (5)
- Modeling and simulation (5)
Institute
- Fakultät Elektrotechnik, Medizintechnik und Informatik (EMI) (ab 04/2019) (464) (remove)
Open Access
- Open Access (233)
- Closed Access (136)
- Closed (83)
- Bronze (52)
- Gold (28)
- Diamond (24)
- Grün (3)
- Hybrid (3)
Verfahren zum Betrieb eines batterieelektrischen Fahrzeugs mit einer elektrischen Maschine zum Antrieb des Fahrzeugs und einem Inverter (1) zum Ansteuern der elektrischen Maschine, wobei der Inverter (1) eine dreiphasige Brückenschaltung mit einer Anzahl von als Halbleiter ausgebildeten Schaltern (3) umfasst, wobei im Inverter (1) entstehende Verluste zum Heizen eines Innenraums des Fahrzeugs und/oder zum Temperieren einer Batterie und/oder zum Temperieren von Getriebeöl verwendet werden, wobei der Inverter (1) mittels Raumzeigermodulation gesteuert wird, wobei ein nicht-optimales Schaltverhalten des Inverters (1) herbeigeführt wird, indem nicht optimale Spannungs-Raumzeiger (e, eu, ev, ew, e1, e2, -e1, -e2) eingestellt werden, wobei eine Skalierung der Spannungs-Raumzeiger (e, e1, e2) über die Schaltung von Nullspannungsvektoren, die je nach zeitlichem Anteil die Spannung reduzieren, oder durch Zuhilfenahme eines jeweils gegenüberliegenden Spannungs-Raumzeigers (-e1, -e2) erfolgt, so dass eine Schaltfolge mit einer maximalen Anzahl von Schaltzyklen realisiert wird, dadurch gekennzeichnet, dass in der Mitte einer Schaltperiode (Tp) keine Symmetrie erzeugt wird.
An in-depth study of U-net for seismic data conditioning: Multiple removal by moveout discrimination
(2024)
Seismic processing often involves suppressing multiples that are an inherent component of collected seismic data. Elaborate multiple prediction and subtraction schemes such as surface-related multiple removal have become standard in industry workflows. In cases of limited spatial sampling, low signal-to-noise ratio, or conservative subtraction of the predicted multiples, the processed data frequently suffer from residual multiples. To tackle these artifacts in the postmigration domain, practitioners often rely on Radon transform-based algorithms. However, such traditional approaches are both time-consuming and parameter dependent, making them relatively complex. In this work, we present a deep learning-based alternative that provides competitive results, while reducing the complexity of its usage, and, hence simplifying its applicability. Our proposed model demonstrates excellent performance when applied to complex field data, despite it being exclusively trained on synthetic data. Furthermore, extensive experiments show that our method can preserve the inherent characteristics of the data, avoiding undesired oversmoothed results, while removing the multiples from seismic offset or angle gathers. Finally, we conduct an in-depth analysis of the model, where we pinpoint the effects of the main hyperparameters on real data inference, and we probabilistically assess its performance from a Bayesian perspective. In this study, we put particular emphasis on helping the user reveal the inner workings of the neural network and attempt to unbox the model.
Analysing and predicting the advance rate of a tunnel boring machine (TBM) in hard rock is integral to tunnelling project planning and execution. It has been applied in the industry for several decades with varying success. Most prediction models are based on or designed for large-diameter TBMs, and much research has been conducted on related tunnelling projects. However, only a few models incorporate information from projects with an outer diameter smaller than 5 m and no penetration prediction model for pipe jacking machines exists to date. In contrast to large TBMs, small-diameter TBMs and their projects have been considered little in research. In general, they are characterised by distinctive features, including insufficient geotechnical information, sometimes rather short drive lengths, special machine designs and partially concurring lining methods like pipe jacking and segment lining. A database which covers most of the parameters mentioned above has been compiled to investigate the performance of small-diameter TBMs in hard rock. In order to provide sufficient geological and technical variance, this database contains 37 projects with 70 geotechnically homogeneous areas. Besides the technical parameters, important geotechnical data like lithological information, unconfined compressive strength, tensile strength and point load index is included and evaluated. The analysis shows that segment lining TBMs have considerably higher penetration rates in similar geological and technical settings mostly due to their design parameters. Different methodologies for predicting TBM penetration, including state-of-the-art models from the literature as well as newly derived regression and machine learning models, are discussed and deployed for backward modelling of the projects contained in the database. New ranges of application for small-diameter tunnelling in several industry-standard penetration models are presented, and new approaches for the penetration prediction of pipe jacking machines in hard rock are proposed.
Physical unclonable functions (PUFs) are increasingly generating attention in the field of hardware-based security for the Internet of Things (IoT). A PUF, as its name implies, is a physical element with a special and unique inherent characteristic and can act as the security anchor for authentication and cryptographic applications. Keeping in mind that the PUF outputs are prone to change in the presence of noise and environmental variations, it is critical to derive reliable keys from the PUF and to use the maximum entropy at the same time. In this work, the PUF output positioning (POP) method is proposed, which is a novel method for grouping the PUF outputs in order to maximize the extracted entropy. To achieve this, an offset data is introduced as helper data, which is used to relax the constraints considered for the grouping of PUF outputs, and deriving more entropy, while reducing the secret key error bits. To implement the method, the key enrollment and key generation algorithms are presented. Based on a theoretical analysis of the achieved entropy, it is proven that POP can maximize the achieved entropy, while respecting the constraints induced to guarantee the reliability of the secret key. Moreover, a detailed security analysis is presented, which shows the resilience of the method against cyber-security attacks. The findings of this work are evaluated by applying the method on a hybrid printed PUF, where it can be practically shown that the proposed method outperforms other existing group-based PUF key generation methods.
This paper provides a comprehensive overview of approaches to the determination of isocontours and isosurfaces from given data sets. Different algorithms are reported in the literature for this purpose, which originate from various application areas, such as computer graphics or medical imaging procedures. In all these applications, the challenge is to extract surfaces with a specific isovalue from a given characteristic, so called isosurfaces. These different application areas have given rise to solution approaches that all solve the problem of isocontouring in their own way. Based on the literature, the following four dominant methods can be identified: the marching cubes algorithms, the tessellation-based algorithms, the surface nets algorithms and the ray tracing algorithms. With regard to their application, it can be seen that the methods are mainly used in the fields of medical imaging, computer graphics and the visualization of simulation results. In our work, we provide a broad and compact overview of the common methods that are currently used in terms of isocontouring with respect to certain criteria and their individual limitations. In this context, we discuss the individual methods and identify possible future research directions in the field of isocontouring.
With the expansion of IoT devices in many aspects of our life, the security of such systems has become an important challenge. Unlike conventional computer systems, any IoT security solution should consider the constraints of these systems such as computational capability, memory, connectivity, and power consumption limitations. Physical Unclonable Functions (PUFs) with their special characteristics were introduced to satisfy the security needs while respecting the mentioned constraints. They exploit the uncontrollable and reproducible variations of the underlying component for security applications such as identification, authentication, and communication security. Since IoT devices are typically low cost, it is important to reuse existing elements in their hardware (for instance sensors, ADCs, etc.) instead of adding extra costs for the PUF hardware. Micro-electromechanical system (MEMS) devices are widely used in IoT systems as sensors and actuators. In this thesis, a comprehensive study of the potential application of MEMS devices as PUF primitives is provided. MEMS PUF leverages the uncontrollable variations in the parameters of MEMS elements to derive secure keys for cryptographic applications. Experimental and simulation results show that our proposed MEMS PUFs are capable of generating enough entropy for a complex key generation, while their responses show low fluctuations in different environmental conditions.
Keeping in mind that the PUF responses are prone to change in the presence of noise and environmental variations, it is critical to derive reliable keys from the PUF and to use the maximum entropy at the same time. In the second part of this thesis, we elaborate on different key generation schemes and their advantages and drawbacks. We propose the PUF output positioning (POP) and integer linear programming (ILP) methods, which are novel methods for grouping the PUF outputs in order to maximize the extracted entropy. To implement these methods, the key enrollment and key generation algorithms are presented. The proposed methods are then evaluated by applying on the responses of the MEMS PUF, where it can be practically shown that the proposed method outperforms other existing PUF key generation methods.
The final part of this thesis is dedicated to the application of the MEMS PUF as a security solution for IoT systems. We select the mutual authentication of IoT devices and their backend system, and propose two lightweight authentication protocols based on MEMS PUFs. The presented protocols undergo a comprehensive security analysis to show their eligibility to be used in IoT systems. As the result, the output of this thesis is a lightweight security solution based on MEMS PUFs, which introduces a very low overhead on the cost of the hardware.
Ultra-low-power passive telemetry systems for industrial and biomedical applications have gained much popularity lately. The reduction of the power consumption and size of the circuits poses critical challenges in ultra-low-power circuit design. Biotelemetry applications like leakage detection in silicone breast implants require low-power-consuming small-size electronics. In this doctoral thesis, the design, simulation, and measurement of a programmable mixed-signal System-on-Chip (SoC) called General Application Passive Sensor Integrated Circuit (GAPSIC) is presented. Owing to the low power consumption, GAPSIC is capable of completely passive operation. Such a batteryless passive system has lower maintenance complexity and is also free from battery-related health hazards. With a die area of 4.92 mm² and a maximum analog power consumption of 592 µW, GAPSIC has one of the best figure-of-merits compared to similar state-of-the-art SoCs. Regarding possible applications, GAPSIC can read out and digitally transmit the signals of resistive sensors for pressure or temperature measurements. Additionally, GAPSIC can measure electrocardiogram (ECG) signals and conductivity.
The design of GAPSIC complies with the International Organization for Standardization (ISO) 15693/NFC (near field communication) 5 standard for radio frequency identification (RFID), corresponding to the frequency range of 13.56 MHz. A passive transponder developed with GAPSIC comprises of an external memory storage and very few other external components, like an antenna and sensors. The passive tag antenna and reader antenna use inductive coupling for communication and energy transfer, which enables passive operation. A passive tag developed with GAPSIC can communicate with an NFC compatible smart device or an ISO 15693 RFID reader. An external memory storage contains the programmable application-specific firmware.
As a mixed-signal SoC, GAPSIC includes both analog and digital circuitries. The analog block of GAPSIC includes a power management unit, an RFID/NFC communication unit, and a sensor readout unit. The digital block includes an integrated 32-bit microcontroller, developed by the Hochschule Offenburg ASIC design center, and digital peripherals. A 16-kilobyte random-access memory and a read-only 16-kilobyte memory constitute the GAPSIC internal memory. For the fabrication of GAPSIC, one poly, six-metal 0.18 µm CMOS process is used.
The design of GAPSIC includes two stages. In the first stage, a standalone RFID/NFC frontend chip with a power management unit, an RFID/NFC communication unit, a clock regenerator unit, and a field detector unit was designed. In the second stage, the rest of the functional blocks were integrated with the blocks of the RFID/NFC frontend chip for the final integration of GAPSIC. To reduce the power consumption, conventional low-power design techniques were applied extensively like multiple power supplies, and the operation of complementary metal-oxide-semiconductor (CMOS) transistors in the sub-threshold region of operation, as well as further innovative circuit designs.
An overvoltage protection circuit, a power rectifier, a bandgap reference circuit, and two low-dropout (LDO) voltage regulators constitute the power management unit of GAPSIC. The overvoltage protection circuit uses a novel method where three stacked transistor pairs shunt the extra voltage. In the power rectifier, four rectifier units are arranged in parallel, which is a unique approach. The four parallel rectifier units provide the optimal choice in terms of voltage drop and the area required.
The communication unit is responsible for RFID/NFC communication and incorporates demodulation and load modulation circuitry. The demodulator circuit comprises of an envelope detector, a high-pass filter, and a comparator. Following a new approach, the bandgap reference circuit itself acts as the load for the envelope detector circuit, which minimizes the circuit complexity and area. For the communication between the reader and the RFID/NFC tag, amplitude-shift keying (ASK) is used to modulate signals, where the smallest modulation index can be as low as 10%. A novel technique involving a comparator with a preset offset voltage effectively demodulates the ASK signal. With an effective die area of 0.7 mm² and power consumption of 107 µW, the standalone RFID/NFC frontend chip has the best figure-of-merits compared to the state-of-the-art frontend chips reported in the relevant literature. A passive RFID/NFC tag developed with the standalone frontend chip, as well as temperature and pressure sensors demonstrate the full passive operational capability of the frontend chip. An NFC reader device using a custom-built Android-based application software reads out the sensor data from the passive tag.
The sensor readout circuit consists of a channel selector with two differential and four single-ended inputs with a programmable-gain instrumentation amplifier. The entire sensor readout part remains deactivated when not in use. The internal memory stores the measured offset voltage of the instrumentation amplifier, where a firmware code removes the offset voltage from the measured sensor signal. A 12-bit successive approximation register (SAR) type analog-to-digital-converter (ADC) based on a charge redistribution architecture converts the measured sensor data to a digital value. The digital peripherals include a serial peripheral interface, four timers, RFID/NFC interfaces, sensor readout unit interfaces, and 12-bit SAR logic.
Two sets of studies with custom-made NFC tag antennas for biomedical applications were conducted to ascertain their compatibility with GAPSIC. The first study involved the link efficiency measurements of NFC tag antennas and an NFC reader antenna with porcine tissue. In a separate experiment, the effect of a ferrite compared to air core on the antenna-coupling factor was investigated. With the ferrite core, the coupling factor increased by four times.
Among the state-of-the-art SoCs published in recent scientific articles, GAPSIC is the only passive programmable SoC with a power management unit, an RFID/NFC communication interface, a sensor readout circuit, a 12-bit SAR ADC, and an integrated 32-bit microcontroller. This doctoral research includes the preliminary study of three passive RFID tags designed with discrete components for biomedical and industrial applications like measurements of temperature, pH, conductivity, and oxygen concentration, along with leakage detection in silicone breast implants. Besides its small size and low power consumption, GAPSIC is suitable for each of the biomedical and industrial applications mentioned above due to the integrated high-performance microcontroller, the robust programmable instrumentation amplifier, and the 12-bit analog-to-digital converter. Furthermore, the simulation and measurement data show that GAPSIC is well suited for the design of a passive tag to monitor arterial blood pressure in patients experiencing Peripheral Artery Disease (PAD), which is proposed in this doctoral thesis as an exemplary application of the developed system.
In many application areas, Deep Reinforcement Learning (DRL) has led to breakthroughs. In Curriculum Learning, the Machine Learning algorithm is not randomly presented with examples, but in a meaningful order of increasing difficulty. This has been used in many application areas to further improve the results of learning systems or to reduce their learning time. Such approaches range from learning plans created manually by domain experts to those created automatically. The automated creation of learning plans is one of the biggest challenges.In this work, we investigate an approach in which a trainer learns in parallel and analogously to the student to automatically create a learning plan for the student during this Double Deep Reinforcement Learning (DDRL). Three Reward functions, Friendly, Adversarial, and Dynamic based on the learner’s reward are compared. The domain for evaluation is kicking with variable distance, direction and relative ball position in the SimSpark simulated soccer environment.As a result, Statistic Curriculum Learning (SCL) performs better than a random curriculum with respect to training time and result quality. DDRL reaches a comparable quality as the baseline and outperforms it significantly in shorter trainings in the distance-direction subdomain reducing the number of required training cycles by almost 50%.
The mathematical representations of data in the Spherical Harmonic (SH) domain has recently regained increasing interest in the machine learning community. This technical report gives an in-depth introduction to the theoretical foundation and practical implementation of SH representations, summarizing works on rotation invariant and equivariant features, as well as convolutions and exact correlations of signals on spheres. In extension, these methods are then generalized from scalar SH representations to Vectorial Harmonics (VH), providing the same capabilities for 3d vector fields on spheres.
Team description papers of magmaOffenburg are incremental in the sense that each year we address a different topic of our team and the tools around our team. In this year’s team description paper we focus on the architecture of the software. It is a main factor for being able to keep the code maintainable even after 15 years of development. We also describe how we make sure that the code follows this architecture.
Ensuring that software applications present their users the most recent version of data is not trivial. Self-adjusting computations are a technique for automatically and efficiently recomputing output data whenever some input changes.
This article describes the software architecture of a large, commercial software system built around a framework for coarse-grained self-adjusting computations in Haskell. It discusses advantages and disadvantages based on longtime experience. The article also presents a demo of the system and explains the API of the framework.
Assessing the robustness of deep neural networks against out-of-distribution inputs is crucial, especially in safety-critical domains like autonomous driving, but also in safety systems where malicious actors can digitally alter inputs to circumvent safety guards. However, designing effective out-of-distribution tests that encompass all possible scenarios while preserving accurate label information is a challenging task. Existing methodologies often entail a compromise between variety and constraint levels for attacks and sometimes even both. In a first step towards a more holistic robustness evaluation of image classification models, we introduce an attack method based on image solarization that is conceptually straightforward yet avoids jeopardizing the global structure of natural images independent of the intensity. Through comprehensive evaluations of multiple ImageNet models, we demonstrate the attack's capacity to degrade accuracy significantly, provided it is not integrated into the training augmentations. Interestingly, even then, no full immunity to accuracy deterioration is achieved. In other settings, the attack can often be simplified into a black-box attack with model-independent parameters. Defenses against other corruptions do not consistently extend to be effective against our specific attack.
Project website: https://github.com/paulgavrikov/adversarial_solarization
Fix your downsampling ASAP! Be natively more robust via Aliasing and Spectral Artifact free Pooling
(2023)
Convolutional neural networks encode images through a sequence of convolutions, normalizations and non-linearities as well as downsampling operations into potentially strong semantic embeddings. Yet, previous work showed that even slight mistakes during sampling, leading to aliasing, can be directly attributed to the networks' lack in robustness. To address such issues and facilitate simpler and faster adversarial training, [12] recently proposed FLC pooling, a method for provably alias-free downsampling - in theory. In this work, we conduct a further analysis through the lens of signal processing and find that such current pooling methods, which address aliasing in the frequency domain, are still prone to spectral leakage artifacts. Hence, we propose aliasing and spectral artifact-free pooling, short ASAP. While only introducing a few modifications to FLC pooling, networks using ASAP as downsampling method exhibit higher native robustness against common corruptions, a property that FLC pooling was missing. ASAP also increases native robustness against adversarial attacks on high and low resolution data while maintaining similar clean accuracy or even outperforming the baseline.
Detecting Images Generated by Deep Diffusion Models using their Local Intrinsic Dimensionality
(2023)
Diffusion models recently have been successfully applied for the visual synthesis of strikingly realistic appearing images. This raises strong concerns about their potential for malicious purposes. In this paper, we propose using the lightweight multi Local Intrinsic Dimensionality (multiLID), which has been originally developed in context of the detection of adversarial examples, for the automatic detection of synthetic images and the identification of the according generator networks. In contrast to many existing detection approaches, which often only work for GAN-generated images, the proposed method provides close to perfect detection results in many realistic use cases. Extensive experiments on known and newly created datasets demonstrate that the proposed multiLID approach exhibits superiority in diffusion detection and model identification.Since the empirical evaluations of recent publications on the detection of generated images are often mainly focused on the "LSUN-Bedroom" dataset, we further establish a comprehensive benchmark for the detection of diffusion-generated images, including samples from several diffusion models with different image sizes.The code for our experiments is provided at https://github.com/deepfake-study/deepfake-multiLID.
Erlang is a functional programming language with dynamic typing. The language offers great flexibility for destructing values through pattern matching and dynamic type tests. Erlang also comes with a type language supporting parametric polymorphism, equi-recursive types, as well as union and a limited form of intersection types. However, type signatures only serve as documentation; there is no check that a function body conforms to its signature.
Set-theoretic types and semantic subtyping fit Erlang’s feature set very well. They allow expressing nearly all constructs of its type language and provide means for statically checking type signatures. This article brings set-theoretic types to Erlang and demonstrates how existing Erlang code can be statically type checked without or with only minor modifications to the code. Further, the article formalizes the main ingredients of the type system in a small core calculus, reports on an implementation of the system, and compares it with other static type checkers for Erlang.
In recent years, predictive maintenance tasks, especially for bearings, have become increasingly important. Solutions for these use cases concentrate on the classification of faults and the estimation of the Remaining Useful Life (RUL). As of today, these solutions suffer from a lack of training samples. In addition, these solutions often require high-frequency accelerometers, incurring significant costs. To overcome these challenges, this research proposes a combined classification and RUL estimation solution based on a Convolutional Neural Network (CNN) and a Long Short-Term Memory (LSTM) network. This solution relies on a hybrid feature extraction approach, making it especially appropriate for low-cost accelerometers with low sampling frequencies. In addition, it uses transfer learning to be suitable for applications with only a few training samples.
With the rising necessity of explainable artificial intelligence (XAI), we see an increase in task-dependent XAI methods on varying abstraction levels. XAI techniques on a global level explain model behavior and on a local level explain sample predictions. We propose a visual analytics workflow to support seamless transitions between global and local explanations, focusing on attributions and counterfactuals on time series classification. In particular, we adapt local XAI techniques (attributions) that are developed for traditional datasets (images, text) to analyze time series classification, a data type that is typically less intelligible to humans. To generate a global overview, we apply local attribution methods to the data, creating explanations for the whole dataset. These explanations are projected onto two dimensions, depicting model behavior trends, strategies, and decision boundaries. To further inspect the model decision-making as well as potential data errors, a what-if analysis facilitates hypothesis generation and verification on both the global and local levels. We constantly collected and incorporated expert user feedback, as well as insights based on their domain knowledge, resulting in a tailored analysis workflow and system that tightly integrates time series transformations into explanations. Lastly, we present three use cases, verifying that our technique enables users to (1)~explore data transformations and feature relevance, (2)~identify model behavior and decision boundaries, as well as, (3)~the reason for misclassifications.
Entity Matching (EM) defines the task of learning to group objects by transferring semantic concepts from example groups (=entities) to unseen data. Despite the general availability of image data in the context of many EM-problems, most currently available EM-algorithms solely rely on (textual) meta data. In this paper, we introduce the first publicly available large-scale dataset for "visual entity matching", based on a production level use case in the retail domain. Using scanned advertisement leaflets, collected over several years from different European retailers, we provide a total of ~786k manually annotated, high resolution product images containing ~18k different individual retail products which are grouped into ~3k entities. The annotation of these product entities is based on a price comparison task, where each entity forms an equivalence class of comparable products. Following on a first baseline evaluation, we show that the proposed "visual entity matching" constitutes a novel learning problem which can not sufficiently be solved using standard image based classification and retrieval algorithms. Instead, novel approaches which allow to transfer example based visual equivalent classes to new data are needed to address the proposed problem. The aim of this paper is to provide a benchmark for such algorithms.
Information about the dataset, evaluation code and download instructions are provided under https://www.retail-786k.org/.
Following the traditional paradigm of convolutional neural networks (CNNs), modern CNNs manage to keep pace with more recent, for example transformer-based, models by not only increasing model depth and width but also the kernel size. This results in large amounts of learnable model parameters that need to be handled during training. While following the convolutional paradigm with the according spatial inductive bias, we question the significance of \emph{learned} convolution filters. In fact, our findings demonstrate that many contemporary CNN architectures can achieve high test accuracies without ever updating randomly initialized (spatial) convolution filters. Instead, simple linear combinations (implemented through efficient 1×1 convolutions) suffice to effectively recombine even random filters into expressive network operators. Furthermore, these combinations of random filters can implicitly regularize the resulting operations, mitigating overfitting and enhancing overall performance and robustness. Conversely, retaining the ability to learn filter updates can impair network performance. Lastly, although we only observe relatively small gains from learning 3×3 convolutions, the learning gains increase proportionally with kernel size, owing to the non-idealities of the independent and identically distributed (\textit{i.i.d.}) nature of default initialization techniques.
Modern CNNs are learning the weights of vast numbers of convolutional operators. In this paper, we raise the fundamental question if this is actually necessary. We show that even in the extreme case of only randomly initializing and never updating spatial filters, certain CNN architectures can be trained to surpass the accuracy of standard training. By reinterpreting the notion of pointwise ($1\times 1$) convolutions as an operator to learn linear combinations (LC) of frozen (random) spatial filters, we are able to analyze these effects and propose a generic LC convolution block that allows tuning of the linear combination rate. Empirically, we show that this approach not only allows us to reach high test accuracies on CIFAR and ImageNet but also has favorable properties regarding model robustness, generalization, sparsity, and the total number of necessary weights. Additionally, we propose a novel weight sharing mechanism, which allows sharing of a single weight tensor between all spatial convolution layers to massively reduce the number of weights.