-
Unintended Bias in 2D+ Image Segmentation and Its Effect on Attention Asymmetry
Authors:
Zsófia Molnár,
Gergely Szabó,
András Horváth
Abstract:
Supervised pretrained models have become widely used in deep learning, especially for image segmentation tasks. However, when applied to specialized datasets such as biomedical imaging, pretrained weights often introduce unintended biases. These biases cause models to assign different levels of importance to different slices, leading to inconsistencies in feature utilization, which can be observed…
▽ More
Supervised pretrained models have become widely used in deep learning, especially for image segmentation tasks. However, when applied to specialized datasets such as biomedical imaging, pretrained weights often introduce unintended biases. These biases cause models to assign different levels of importance to different slices, leading to inconsistencies in feature utilization, which can be observed as asymmetries in saliency map distributions. This transfer of color distributions from natural images to non-natural datasets can compromise model performance and reduce the reliability of results. In this study, we investigate the effects of these biases and propose strategies to mitigate them. Through a series of experiments, we test both pretrained and randomly initialized models, comparing their performance and saliency map distributions. Our proposed methods, which aim to neutralize the bias introduced by pretrained color channel weights, demonstrate promising results, offering a practical approach to improving model explainability while maintaining the benefits of pretrained models. This publication presents our findings, providing insights into addressing pretrained weight biases across various deep learning tasks.
△ Less
Submitted 20 May, 2025;
originally announced May 2025.
-
Statistical process discovery
Authors:
Pierre Cry,
Paolo Ballarini,
András Horváth,
Pascale Le Gall
Abstract:
Stochastic process discovery is concerned with deriving a model capable of reproducing the stochastic character of observed executions of a given process, stored in a log. This leads to an optimisation problem in which the model's parameter space is searched for, driven by the resemblance between the log's and the model's stochastic languages. The bottleneck of such optimisation problem lay in the…
▽ More
Stochastic process discovery is concerned with deriving a model capable of reproducing the stochastic character of observed executions of a given process, stored in a log. This leads to an optimisation problem in which the model's parameter space is searched for, driven by the resemblance between the log's and the model's stochastic languages. The bottleneck of such optimisation problem lay in the determination of the model's stochastic language which existing approaches deal with through, hardly scalable, exact computation approaches. In this paper we introduce a novel framework in which we combine a simulation-based Bayesian parameter inference scheme, used to search for the ``optimal'' instance of a stochastic model, with an expressive statistical model checking engine, used (during inference) to approximate the language of the considered model's instance. Because of its simulation-based nature, the payoff is that, the runtime for discovering of the optimal instance of a model can be easily traded in for accuracy, hence allowing to treat large models which would result in a prohibitive runtime with non-simulation based alternatives. We validate our approach on several popular event logs concerning real-life systems.
△ Less
Submitted 30 April, 2025;
originally announced April 2025.
-
Bio-crafting Architecture: Experiences of growing mycelium in minimal surface molds
Authors:
Anca-Simona Horvath,
Alina Elena Voinea,
Radu Arieşan
Abstract:
This study documents a three-week workshop with architecture students, where we designed and 3D printed various minimal surfaces using wood-based filaments, and used them as molds in which to grow mycelium. We detail the design process and the growth of the mycelium in different shapes, together with participants' experiences of working with a living material. After exhibiting the results of the w…
▽ More
This study documents a three-week workshop with architecture students, where we designed and 3D printed various minimal surfaces using wood-based filaments, and used them as molds in which to grow mycelium. We detail the design process and the growth of the mycelium in different shapes, together with participants' experiences of working with a living material. After exhibiting the results of the work in a public-facing exhibition, we conducted interviews with members of the general public about their perceptions on interacting with a material such as mycelium in design. Our findings show that 3D-printed minimal surfaces with wood-based filaments can function as structural cores for mycelium-based composites and mycelium binds to the filament. Participants in the workshop exhibited stronger feelings for living materials compared to non-living ones, displaying both biophilia and, to a lesser extent, biophobia when interacting with the mycelium. Members of the general public discuss pragmatic aspects including mold, fragility, or production costs, and speculate on the future of bio-technology and its impact on everyday life. While all are positive about the impact on bio-technologies on the future, they have diverging opinions on how much ethical considerations should influence research directions.
△ Less
Submitted 21 March, 2025;
originally announced April 2025.
-
Removable Sets for Fractional Heat and Fractional Bessel-Heat Equations
Authors:
Mouna Chegaar,
Á. P. Horváth
Abstract:
We examine the fractional heat diffusion equations $L_{γ,a}:=(-Δ_a)^{\fracγ{2}}+\partial_t$, where $Δ_a$ is the Laplace- or the Bessel-Laplace operator. We give conditions for removability which are sufficient and which are necessary, by $L^p$-capacities. Introducing a spherical modulus of smoothness we can treat the Laplace and Bessel-Laplace cases together.
We examine the fractional heat diffusion equations $L_{γ,a}:=(-Δ_a)^{\fracγ{2}}+\partial_t$, where $Δ_a$ is the Laplace- or the Bessel-Laplace operator. We give conditions for removability which are sufficient and which are necessary, by $L^p$-capacities. Introducing a spherical modulus of smoothness we can treat the Laplace and Bessel-Laplace cases together.
△ Less
Submitted 13 April, 2025;
originally announced April 2025.
-
Probabilistic Process Discovery with Stochastic Process Trees
Authors:
András Horváth,
Paolo Ballarini,
Pierre Cry
Abstract:
In order to obtain a stochastic model that accounts for the stochastic aspects of the dynamics of a business process, usually the following steps are taken. Given an event log, a process tree is obtained through a process discovery algorithm, i.e., a process tree that is aimed at reproducing, as accurately as possible, the language of the log. The process tree is then transformed into a Petri net…
▽ More
In order to obtain a stochastic model that accounts for the stochastic aspects of the dynamics of a business process, usually the following steps are taken. Given an event log, a process tree is obtained through a process discovery algorithm, i.e., a process tree that is aimed at reproducing, as accurately as possible, the language of the log. The process tree is then transformed into a Petri net that generates the same set of sequences as the process tree. In order to capture the frequency of the sequences in the event log, weights are assigned to the transitions of the Petri net, resulting in a stochastic Petri net with a stochastic language in which each sequence is associated with a probability. In this paper we show that this procedure has unfavorable properties. First, the weights assigned to the transitions of the Petri net have an unclear role in the resulting stochastic language. We will show that a weight can have multiple, ambiguous impact on the probability of the sequences generated by the Petri net. Second, a number of different Petri nets with different number of transitions can correspond to the same process tree. This means that the number of parameters (the number of weights) that determines the stochastic language is not well-defined. In order to avoid these ambiguities, in this paper, we propose to add stochasticity directly to process trees. The result is a new formalism, called stochastic process trees, in which the number of parameters and their role in the associated stochastic language is clear and well-defined.
△ Less
Submitted 8 April, 2025;
originally announced April 2025.
-
Edge Training and Inference with Analog ReRAM Technology for Hand Gesture Recognition
Authors:
Victoria Clerico,
Anirvan Dutta,
Donato Francesco Falcone,
Wooseok Choi,
Matteo Galetta,
Tommaso Stecconi,
András Horváth,
Shokoofeh Varzandeh,
Bert Jan Offrein,
Mohsen Kaboli,
Valeria Bragaglia
Abstract:
Tactile hand gesture recognition is a crucial task for user control in the automotive sector, where Human-Machine Interactions (HMI) demand low latency and high energy efficiency. This study addresses the challenges of power-constrained edge training and inference by utilizing analog Resistive Random Access Memory (ReRAM) technology in conjunction with a real tactile hand gesture dataset. By optim…
▽ More
Tactile hand gesture recognition is a crucial task for user control in the automotive sector, where Human-Machine Interactions (HMI) demand low latency and high energy efficiency. This study addresses the challenges of power-constrained edge training and inference by utilizing analog Resistive Random Access Memory (ReRAM) technology in conjunction with a real tactile hand gesture dataset. By optimizing the input space through a feature engineering strategy, we avoid relying on large-scale crossbar arrays, making the system more suitable for edge deployment. Through realistic hardware-aware simulations that account for device non-idealities derived from experimental data, we demonstrate the functionalities of our analog ReRAM-based analog in-memory computing for on-chip training, utilizing the state-of-the-art Tiki-Taka algorithm. Furthermore, we validate the classification accuracy of approximately 91.4% for post-deployment inference of hand gestures. The results highlight the potential of analog ReRAM technology and crossbar architecture with fully parallelized matrix computations for real-time HMI systems at the Edge.
△ Less
Submitted 25 February, 2025;
originally announced February 2025.
-
Speed of sound in Kaluza-Klein Fermi gas
Authors:
Anna Horváth,
Emese Forgács-Dajka,
Gergely Gábor Barnaföldi
Abstract:
A five-dimensional Kaluza-Klein spacetime model is considered, with one extra compactified spatial dimension. The equation of state of an electrically neutral, zero-temperature Fermi gas with a repulsive linear potential is described. From the equation of state, the speed of sound squared is calculated and shown for different model parameters. Its properties are studied from lower energies up to t…
▽ More
A five-dimensional Kaluza-Klein spacetime model is considered, with one extra compactified spatial dimension. The equation of state of an electrically neutral, zero-temperature Fermi gas with a repulsive linear potential is described. From the equation of state, the speed of sound squared is calculated and shown for different model parameters. Its properties are studied from lower energies up to the conformal limit.
△ Less
Submitted 7 February, 2025;
originally announced February 2025.
-
Strongly self-dual polytopes
Authors:
Ákos G. Horváth,
István Prok
Abstract:
This article aims to study the class of strongly self-dual polytopes (ssd-polytopes for short), defined in a paper by Lovász \cite{lovasz}. He described a series of such polytopes (called $L$-type polytopes), which he used to solve a combinatorial problem. From a geometrical point of view, there are interesting questions: what additional elements of this class exist, and are there any with a diffe…
▽ More
This article aims to study the class of strongly self-dual polytopes (ssd-polytopes for short), defined in a paper by Lovász \cite{lovasz}. He described a series of such polytopes (called $L$-type polytopes), which he used to solve a combinatorial problem. From a geometrical point of view, there are interesting questions: what additional elements of this class exist, and are there any with a different structure from the $L$-type ones? We show that in dimension three, one of their faces defines $L$-type polyhedra. Illustrating the algorithm of the proof, we present an ssd-polytope of 23 vertices whose combinatorial structure differ from those of $L$-type ones. Finally, with an elementary discussion, we prove that for fewer than nine vertices, there are only fifth one ssd-polyhedra, four of them can be constructed by Lovász's method, and we can find the fifth one with "the diameter gradient flow algorithm" of Katz, Memoli and Wang \cite{katz-memoli-wang}.
△ Less
Submitted 27 January, 2025;
originally announced January 2025.
-
Post-Hoc MOTS: Exploring the Capabilities of Time-Symmetric Multi-Object Tracking
Authors:
Gergely Szabó,
Zsófia Molnár,
András Horváth
Abstract:
Temporal forward-tracking has been the dominant approach for multi-object segmentation and tracking (MOTS). However, a novel time-symmetric tracking methodology has recently been introduced for the detection, segmentation, and tracking of budding yeast cells in pre-recorded samples. Although this architecture has demonstrated a unique perspective on stable and consistent tracking, as well as misse…
▽ More
Temporal forward-tracking has been the dominant approach for multi-object segmentation and tracking (MOTS). However, a novel time-symmetric tracking methodology has recently been introduced for the detection, segmentation, and tracking of budding yeast cells in pre-recorded samples. Although this architecture has demonstrated a unique perspective on stable and consistent tracking, as well as missed instance re-interpolation, its evaluation has so far been largely confined to settings related to videomicroscopic environments. In this work, we aim to reveal the broader capabilities, advantages, and potential challenges of this architecture across various specifically designed scenarios, including a pedestrian tracking dataset. We also conduct an ablation study comparing the model against its restricted variants and the widely used Kalman filter. Furthermore, we present an attention analysis of the tracking architecture for both pretrained and non-pretrained models
△ Less
Submitted 11 December, 2024;
originally announced December 2024.
-
Soft cells, Kelvin's foam and the minimal surfaces of Schwarz
Authors:
Gábor Domokos,
Alain Goriely,
Ákos G. Horváth,
Krisztina Regős
Abstract:
Recently, we introduced a new class of shapes, called soft cells which fill space as soft tilings without gaps and overlaps while minimizing the number of sharp corners. We introduced the edge bending algorithm that deforms a polyhedral tiling into a soft tiling and we proved that an infinite class of polyhedral tilings can be smoothly deformed into standard soft tilings. Here, we demonstrate that…
▽ More
Recently, we introduced a new class of shapes, called soft cells which fill space as soft tilings without gaps and overlaps while minimizing the number of sharp corners. We introduced the edge bending algorithm that deforms a polyhedral tiling into a soft tiling and we proved that an infinite class of polyhedral tilings can be smoothly deformed into standard soft tilings. Here, we demonstrate that certain triply periodic minimal surfaces naturally give rise to non-standard soft tilings. By extending the edge-bending algorithm, we further establish that the soft tilings derived from the Schwarz P and Schwarz D surfaces can be continuously transformed into one another through a one-parameter family of intermediate non-standard soft tilings. Notably, by carrying its combinatorial structure, both resulting tilings belong to the first order equivalence class of the Dirichlet-Voronoi tiling on the body-centered cubic bcc lattice, highlighting a deep geometric connection underlying these minimal surface configurations. By requiring identical end-tangents for edges in a first order class, we also define second order equivalence classes among tilings and prove that there exist exactly two such classes among soft tilings which share the full symmetry group of the DV-bcc tiling. Additionally, we construct a one-parameter family of tilings bridging standard and non-standard soft tilings, explicitly including the classic Kelvin foam structure as an intermediate configuration. This construction highlights that both the soft cells themselves and the geometric methods employed in their generation provide valuable insights into the structural principles underlying natural forms. We also present the soft tiling induced by the gyroid structure.
△ Less
Submitted 8 April, 2025; v1 submitted 23 November, 2024;
originally announced December 2024.
-
Stable Diffusion with Continuous-time Neural Network
Authors:
Andras Horvath
Abstract:
Stable diffusion models have ushered in a new era of advancements in image generation, currently reigning as the state-of-the-art approach, exhibiting unparalleled performance. The process of diffusion, accompanied by denoising through iterative convolutional or transformer network steps, stands at the core of their implementation. Neural networks operating in continuous time naturally embrace the…
▽ More
Stable diffusion models have ushered in a new era of advancements in image generation, currently reigning as the state-of-the-art approach, exhibiting unparalleled performance. The process of diffusion, accompanied by denoising through iterative convolutional or transformer network steps, stands at the core of their implementation. Neural networks operating in continuous time naturally embrace the concept of diffusion, this way they could enable more accurate and energy efficient implementation.
Within the confines of this paper, my focus delves into an exploration and demonstration of the potential of celllular neural networks in image generation. I will demonstrate their superiority in performance, showcasing their adeptness in producing higher quality images and achieving quicker training times in comparison to their discrete-time counterparts on the commonly cited MNIST dataset.
△ Less
Submitted 16 October, 2024;
originally announced October 2024.
-
The effect of multiple extra dimensions on the maximal mass of compact stars in Kaluza-Klein space-time
Authors:
Anna Horváth,
Emese Forgács-Dajka,
Gergely Gábor Barnaföldi
Abstract:
Compact stars in the Kaluza-Klein space-time are investigated, with multiple additional compactified spatial dimensions ($d$). Within the extended phenomenological model, a static, spherically symmetric solution is considered, with the equation of state provided by a zero temperature, interacting multi-dimensional Fermi gas. The maximal masses of compact stars are calculated for different model pa…
▽ More
Compact stars in the Kaluza-Klein space-time are investigated, with multiple additional compactified spatial dimensions ($d$). Within the extended phenomenological model, a static, spherically symmetric solution is considered, with the equation of state provided by a zero temperature, interacting multi-dimensional Fermi gas. The maximal masses of compact stars are calculated for different model parameters. We investigated the effect of the existence of multiple extra compactified dimensions within the Kaluza-Klein compact star structure. We found that the number of extra dimensions plays a similar role, and to a similar order, as the excitation number: increasing their number, $d$, reduces the maximal mass by a few percent.
△ Less
Submitted 14 October, 2024;
originally announced October 2024.
-
FRIDA: Free-Rider Detection using Privacy Attacks
Authors:
Pol G. Recasens,
Ádám Horváth,
Alberto Gutierrez-Torre,
Jordi Torres,
Josep Ll. Berral,
Balázs Pejó
Abstract:
Federated learning is increasingly popular as it enables multiple parties with limited datasets and resources to train a high-performing machine learning model collaboratively. However, similarly to other collaborative systems, federated learning is vulnerable to free-riders -- participants who do not contribute to the training but still benefit from the shared model. Free-riders not only compromi…
▽ More
Federated learning is increasingly popular as it enables multiple parties with limited datasets and resources to train a high-performing machine learning model collaboratively. However, similarly to other collaborative systems, federated learning is vulnerable to free-riders -- participants who do not contribute to the training but still benefit from the shared model. Free-riders not only compromise the integrity of the learning process but also slow down the convergence of the global model, resulting in increased costs for the honest participants.
To address this challenge, we propose FRIDA: free-rider detection using privacy attacks, a framework that leverages inference attacks to detect free-riders. Unlike traditional methods that only capture the implicit effects of free-riding, FRIDA directly infers details of the underlying training datasets, revealing characteristics that indicate free-rider behaviour. Through extensive experiments, we demonstrate that membership and property inference attacks are effective for this purpose. Our evaluation shows that FRIDA outperforms state-of-the-art methods, especially in non-IID settings.
△ Less
Submitted 7 October, 2024;
originally announced October 2024.
-
Trace inequality with Bessel convolution
Authors:
Mouna Chegaar,
Á. P. Horváth
Abstract:
Considering potentials defined by Bessel kernel with Bessel convolution a Kerman-Sawyer type characterization of trace inequality is given. As an application an estimate on the least eigenvalue of Schrödinger-Bessel operators is derived.
Considering potentials defined by Bessel kernel with Bessel convolution a Kerman-Sawyer type characterization of trace inequality is given. As an application an estimate on the least eigenvalue of Schrödinger-Bessel operators is derived.
△ Less
Submitted 27 September, 2024;
originally announced September 2024.
-
Application of Kaluza-Klein Theory in Modeling Compact Stars: Exploring Extra Dimensions
Authors:
Anna Horváth,
Emese Forgács-Dajka,
Gergely Gábor Barnaföldi
Abstract:
A theoretical framework for calculating the mass-radius curve of compact stars in the Kaluza-Klein space-time is introduced, with one additional compact spatial dimension. Static, spherically symmetric solutions are considered, with the equation of state provided by a zero temperature, interacting multidimensional Fermi gas. To model the strong force between baryons, a repulsive potential is intro…
▽ More
A theoretical framework for calculating the mass-radius curve of compact stars in the Kaluza-Klein space-time is introduced, with one additional compact spatial dimension. Static, spherically symmetric solutions are considered, with the equation of state provided by a zero temperature, interacting multidimensional Fermi gas. To model the strong force between baryons, a repulsive potential is introduced, which is linear in the particle number density. The maximal mass of compact stars is calculated for different model parameters, and with a physical parameter choice, it satisfies observational data, meaning that it is possible to model simple, realistic objects within this framework. Based on this comparison, a limiting size for the observational regime of extra dimensions in compact stars is provided, with $r_c \gtrsim 0.2$~fm.
△ Less
Submitted 10 September, 2024; v1 submitted 29 August, 2024;
originally announced August 2024.
-
Magicity versus superfluidity around $^{28}$O viewed from the study of $^{30}$F
Authors:
J. Kahlbow,
T. Aumann,
O. Sorlin,
Y. Kondo,
T. Nakamura,
F. Nowacki,
A. Revel,
N. L. Achouri,
H. Al Falou,
L. Atar,
H. Baba,
K. Boretzky,
C. Caesar,
D. Calvet,
H. Chae,
N. Chiga,
A. Corsi,
F. Delaunay,
A. Delbart,
Q. Deshayes,
Z. Dombradi,
C. A. Douma,
Z. Elekes,
I. Gasparic,
J. -M. Gheller
, et al. (62 additional authors not shown)
Abstract:
The neutron-rich unbound fluorine isotope $^{30}$F$_{21}$ has been observed for the first time by measuring its neutron decay at the SAMURAI spectrometer (RIBF, RIKEN) in the quasi-free proton knockout reaction of $^{31}$Ne nuclei at 235 MeV/nucleon. The mass and thus one-neutron-separation energy of $^{30}$F has been determined to be $S_n = -472\pm 58 \mathrm{(stat.)} \pm 33 \mathrm{(sys.)}$ keV…
▽ More
The neutron-rich unbound fluorine isotope $^{30}$F$_{21}$ has been observed for the first time by measuring its neutron decay at the SAMURAI spectrometer (RIBF, RIKEN) in the quasi-free proton knockout reaction of $^{31}$Ne nuclei at 235 MeV/nucleon. The mass and thus one-neutron-separation energy of $^{30}$F has been determined to be $S_n = -472\pm 58 \mathrm{(stat.)} \pm 33 \mathrm{(sys.)}$ keV from the measurement of its invariant-mass spectrum. The absence of a sharp drop in $S_n$($^{30}$F) shows that the ``magic'' $N=20$ shell gap is not restored close to $^{28}$O, which is in agreement with our shell-model calculations that predict a near degeneracy between the neutron $d$ and $fp$ orbitals, with the $1p_{3/2}$ and $1p_{1/2}$ orbitals becoming more bound than the $0f_{7/2}$ one. This degeneracy and reordering of orbitals has two potential consequences: $^{28}$O behaves like a strongly superfluid nucleus with neutron pairs scattering across shells, and both $^{29,31}$F appear to be good two-neutron halo-nucleus candidates.
△ Less
Submitted 27 July, 2024;
originally announced July 2024.
-
A framework for optimisation based stochastic process discovery
Authors:
Pierre Cry,
András Horváth,
Paolo Ballarini,
Pascal Le Gall
Abstract:
Process mining is concerned with deriving formal models capable of reproducing the behaviour of a given organisational process by analysing observed executions collected in an event log. The elements of an event log are finite sequences (i.e., traces or words) of actions. Many effective algorithms have been introduced which issue a control flow model (commonly in Petri net form) aimed at reproduci…
▽ More
Process mining is concerned with deriving formal models capable of reproducing the behaviour of a given organisational process by analysing observed executions collected in an event log. The elements of an event log are finite sequences (i.e., traces or words) of actions. Many effective algorithms have been introduced which issue a control flow model (commonly in Petri net form) aimed at reproducing, as precisely as possible, the language of the considered event log. However, given that identical executions can be observed several times, traces of an event log are associated with a frequency and, hence, an event log inherently yields also a stochastic language. By exploiting the trace frequencies contained in the event log, the stochastic extension of process mining, therefore, consists in deriving stochastic (Petri nets) models capable of reproducing the likelihood of the observed executions. In this paper, we introduce a novel stochastic process mining approach. Starting from a "standard" Petri net model mined through classical mining algorithms, we employ optimization to identify optimal weights for the transitions of the mined net so that the stochastic language issued by the stochastic interpretation of the mined net closely resembles that of the event log. The optimization is either based on the maximum likelihood principle or on the earth moving distance. Experiments on some popular real system logs show an improved accuracy w.r.t. to alternative approaches.
△ Less
Submitted 2 July, 2024; v1 submitted 16 June, 2024;
originally announced June 2024.
-
Soft cells and the geometry of seashells
Authors:
Gábor Domokos,
Alain Goriely,
Ákos G. Horváth,
Krisztina Regős
Abstract:
A central problem of geometry is the tiling of space with simple structures. The classical solutions, such as triangles, squares, and hexagons in the plane and cubes and other polyhedra in three-dimensional space are built with sharp corners and flat faces. However, many tilings in Nature are characterized by shapes with curved edges, non-flat faces, and few, if any, sharp corners. An important qu…
▽ More
A central problem of geometry is the tiling of space with simple structures. The classical solutions, such as triangles, squares, and hexagons in the plane and cubes and other polyhedra in three-dimensional space are built with sharp corners and flat faces. However, many tilings in Nature are characterized by shapes with curved edges, non-flat faces, and few, if any, sharp corners. An important question is then to relate prototypical sharp tilings to softer natural shapes. Here, we solve this problem by introducing a new class of shapes, the \textit{soft cells}, minimizing the number of sharp corners and filling space as \emph{soft tilings}. We prove that an infinite class of polyhedral tilings can be smoothly deformed into soft tilings and we construct the soft versions of all Dirichlet-Voronoi cells associated with point lattices in two and three dimensions. Remarkably, these ideal soft shapes, born out of geometry, are found abundantly in nature, from cells to shells.
△ Less
Submitted 6 February, 2024;
originally announced February 2024.
-
Fold Bifurcation Identification through Scientific Machine Learning
Authors:
Giuseppe Habib,
Ádám Horváth
Abstract:
This study employs scientific machine learning to identify transient time series of dynamical systems near a fold bifurcation of periodic solutions. The unique aspect of this work is that a convolutional neural network (CNN) is trained with a relatively small amount of data and on a single, very simple system, yet it is tested on much more complicated systems. This task requires strong generalizat…
▽ More
This study employs scientific machine learning to identify transient time series of dynamical systems near a fold bifurcation of periodic solutions. The unique aspect of this work is that a convolutional neural network (CNN) is trained with a relatively small amount of data and on a single, very simple system, yet it is tested on much more complicated systems. This task requires strong generalization capabilities, which are achieved by incorporating physics-based information. This information is provided through a specific pre-processing of the input data, which includes transformation into polar coordinates, normalization, transformation into the logarithmic scale, and filtering through a moving mean. The results demonstrate that such data pre-processing enables the CNN to grasp the important features related to transient time-series near a fold bifurcation, namely, the trend of the oscillation amplitude, and disregard other characteristics that are not particularly relevant, such as the vibration frequency. The developed CNN was able to correctly classify transient trajectories near a fold for a mass-on-moving-belt system, a van der Pol-Duffing oscillator with an attached tuned mass damper, and a pitch-and-plunge wing profile. The results contribute to the progress towards the development of similar CNNs effective in real-life applications such as safety monitoring of dynamical systems.
△ Less
Submitted 30 January, 2025; v1 submitted 21 December, 2023;
originally announced December 2023.
-
Elementary Constructions of conic sections
Authors:
Ákos G. Horváth
Abstract:
In classical geometry, there is no such well-known and much-studied topic as the construction of conic sections (or briefly conics) from its five points. Its importance in many applications of mechanical engineering, civil engineering and architectural engineering, as well as other applied sciences is clear. The beauty of the topic is that it raises difficult questions that can be approached with…
▽ More
In classical geometry, there is no such well-known and much-studied topic as the construction of conic sections (or briefly conics) from its five points. Its importance in many applications of mechanical engineering, civil engineering and architectural engineering, as well as other applied sciences is clear. The beauty of the topic is that it raises difficult questions that can be approached with basic tools. In this article, we provide constructions (and corresponding theories) that can be taught to high school and university students without knowledge of projective geometry. For this, we recall some important facts about conic sections that can be found in the rich literature. We use the concepts of power of a point on a circle, similarity, orthogonal affinity and inversion. We also mention famous constructions related to our questions. We begin our article at this point, where the standard teaching ends the discussion of conic sections. We therefore assume that the reader knows the basic definitions and constructions of conics, the concepts of focus, axis, tangent, leading circle and leading line.
△ Less
Submitted 13 October, 2023;
originally announced October 2023.
-
Which model density is best in pair natural orbital local correlation theory?
Authors:
Reka A. Horvath,
Kesha Sorathia,
Isabelle Saint,
David P. Tew
Abstract:
Low-scaling electron correlation theory based on the pair natural orbital approximation, PNO-CCSD(T), has become a powerful computational tool. Motivated by the recent discovery of large errors for organometallic molecules, we assess the role of the model density used to discard unimportant contributions. We find that second-order perturbation theory provides the best compromise between cost and a…
▽ More
Low-scaling electron correlation theory based on the pair natural orbital approximation, PNO-CCSD(T), has become a powerful computational tool. Motivated by the recent discovery of large errors for organometallic molecules, we assess the role of the model density used to discard unimportant contributions. We find that second-order perturbation theory provides the best compromise between cost and accuracy, but coupling between localised occupied orbitals must be accounted for. Errors in the CCSD energy are then well below 1~kcal/mol, even for molecules with moderate multi-reference character, and the primary remaining source of errors lies in the treatment of the (T) energy contribution.
△ Less
Submitted 8 October, 2023;
originally announced October 2023.
-
Targeted Adversarial Attacks on Generalizable Neural Radiance Fields
Authors:
Andras Horvath,
Csaba M. Jozsa
Abstract:
Neural Radiance Fields (NeRFs) have recently emerged as a powerful tool for 3D scene representation and rendering. These data-driven models can learn to synthesize high-quality images from sparse 2D observations, enabling realistic and interactive scene reconstructions. However, the growing usage of NeRFs in critical applications such as augmented reality, robotics, and virtual environments could…
▽ More
Neural Radiance Fields (NeRFs) have recently emerged as a powerful tool for 3D scene representation and rendering. These data-driven models can learn to synthesize high-quality images from sparse 2D observations, enabling realistic and interactive scene reconstructions. However, the growing usage of NeRFs in critical applications such as augmented reality, robotics, and virtual environments could be threatened by adversarial attacks.
In this paper we present how generalizable NeRFs can be attacked by both low-intensity adversarial attacks and adversarial patches, where the later could be robust enough to be used in real world applications. We also demonstrate targeted attacks, where a specific, predefined output scene is generated by these attack with success.
△ Less
Submitted 5 October, 2023;
originally announced October 2023.
-
Enhancing Cell Tracking with a Time-Symmetric Deep Learning Approach
Authors:
Gergely Szabó,
Paolo Bonaiuti,
Andrea Ciliberto,
András Horváth
Abstract:
The accurate tracking of live cells using video microscopy recordings remains a challenging task for popular state-of-the-art image processing based object tracking methods. In recent years, several existing and new applications have attempted to integrate deep-learning based frameworks for this task, but most of them still heavily rely on consecutive frame based tracking embedded in their archite…
▽ More
The accurate tracking of live cells using video microscopy recordings remains a challenging task for popular state-of-the-art image processing based object tracking methods. In recent years, several existing and new applications have attempted to integrate deep-learning based frameworks for this task, but most of them still heavily rely on consecutive frame based tracking embedded in their architecture or other premises that hinder generalized learning. To address this issue, we aimed to develop a new deep-learning based tracking method that relies solely on the assumption that cells can be tracked based on their spatio-temporal neighborhood, without restricting it to consecutive frames. The proposed method has the additional benefit that the motion patterns of the cells can be learned completely by the predictor without any prior assumptions, and it has the potential to handle a large number of video frames with heavy artifacts. The efficacy of the proposed method is demonstrated through biologically motivated validation strategies and compared against multiple state-of-the-art cell tracking methods.
△ Less
Submitted 31 January, 2025; v1 submitted 4 August, 2023;
originally announced August 2023.
-
Intruder configurations in $^{29}$Ne at the transition into the island of inversion: Detailed structure study of $^{28}$Ne
Authors:
H. Wang,
M. Yasuda,
Y. Kondo,
T. Nakamura,
J. A. Tostevin,
K. Ogata,
T. Otsuka,
A. Poves,
N. Shimizu,
K. Yoshida,
N. L. Achouri,
H. Al Falou,
L. Atar,
T. Aumann,
H. Baba,
K. Boretzky,
C. Caesar,
D. Calvet,
H. Chae,
N. Chiga,
A. Corsi,
H. L. Crawford,
F. Delaunay,
A. Delbart,
Q. Deshayes
, et al. (71 additional authors not shown)
Abstract:
Detailed $γ$-ray spectroscopy of the exotic neon isotope $^{28}$Ne has been performed for the first time using the one-neutron removal reaction from $^{29}$Ne on a liquid hydrogen target at 240~MeV/nucleon. Based on an analysis of parallel momentum distributions, a level scheme with spin-parity assignments has been constructed for $^{28}$Ne and the negative-parity states are identified for the fir…
▽ More
Detailed $γ$-ray spectroscopy of the exotic neon isotope $^{28}$Ne has been performed for the first time using the one-neutron removal reaction from $^{29}$Ne on a liquid hydrogen target at 240~MeV/nucleon. Based on an analysis of parallel momentum distributions, a level scheme with spin-parity assignments has been constructed for $^{28}$Ne and the negative-parity states are identified for the first time. The measured partial cross sections and momentum distributions reveal a significant intruder $p$-wave strength providing evidence of the breakdown of the $N=20$ and $N=28$ shell gaps. Only a weak, possible $f$-wave strength was observed to bound final states. Large-scale shell-model calculations with different effective interactions do not reproduce the large $p$-wave and small $f$-wave strength observed experimentally, indicating an ongoing challenge for a complete theoretical description of the transition into the island of inversion along the Ne isotopic chain.
△ Less
Submitted 28 June, 2023;
originally announced June 2023.
-
Event-shape-dependent analysis of charm-anticharm azimuthal correlations in simulations
Authors:
Aniko Horvath,
Eszter Frajna,
Robert Vertesi
Abstract:
In high-energy collisions of small systems, by high-enough final-state multiplicities, a collective behaviour is present that is similar to the flow patterns observed in heavy-ion collisions. Recent studies connect this collectivity to semi-soft vacuum-QCD processes. Here we explore QCD production mechanisms using angular correlations of heavy flavour using simulated proton-proton collisions at…
▽ More
In high-energy collisions of small systems, by high-enough final-state multiplicities, a collective behaviour is present that is similar to the flow patterns observed in heavy-ion collisions. Recent studies connect this collectivity to semi-soft vacuum-QCD processes. Here we explore QCD production mechanisms using angular correlations of heavy flavour using simulated proton-proton collisions at $\sqrt{s} = 13$~TeV with the PYTHIA8 Monte Carlo event generator. We demonstrate that the event shape is strongly connected to the production mechanisms. Flattenicity, a novel event descriptor, can be used to separate events containing the final-state radiation from the rest of the events.
△ Less
Submitted 9 June, 2023;
originally announced June 2023.
-
The bridge between Desargues' and Pappus' theorems
Authors:
Ákos G. Horváth
Abstract:
In this paper, we investigate the configuration theorems of Desargues and Pappus in a synthetic geometric way. We provide a bridge between the two configurations with a third one that can be considered a specification for both. We do not use the theory of collineations or the analytic description of the plane over a ternary ring.
In this paper, we investigate the configuration theorems of Desargues and Pappus in a synthetic geometric way. We provide a bridge between the two configurations with a third one that can be considered a specification for both. We do not use the theory of collineations or the analytic description of the plane over a ternary ring.
△ Less
Submitted 3 May, 2023;
originally announced May 2023.
-
$p$-capacity with Bessel convolution
Authors:
Á. P. Horváth
Abstract:
We define and examine nonlinear potential by Bessel convolution with Bessel kernel. We investigate removable sets with respect to Laplace-Bessel inequality. By studying the maximal and fractional maximal measure, a Wolff type inequality is proved. Finally the relation of B-$p$ capacity and B-Lipschitz mapping, and the B-$p$ capacity and weighted Hausdorff measure and the B-$p$ capacity of Cantor s…
▽ More
We define and examine nonlinear potential by Bessel convolution with Bessel kernel. We investigate removable sets with respect to Laplace-Bessel inequality. By studying the maximal and fractional maximal measure, a Wolff type inequality is proved. Finally the relation of B-$p$ capacity and B-Lipschitz mapping, and the B-$p$ capacity and weighted Hausdorff measure and the B-$p$ capacity of Cantor sets are examined.
△ Less
Submitted 23 March, 2023;
originally announced March 2023.
-
Exploratory Analysis of Federated Learning Methods with Differential Privacy on MIMIC-III
Authors:
Aron N. Horvath,
Matteo Berchier,
Farhad Nooralahzadeh,
Ahmed Allam,
Michael Krauthammer
Abstract:
Background: Federated learning methods offer the possibility of training machine learning models on privacy-sensitive data sets, which cannot be easily shared. Multiple regulations pose strict requirements on the storage and usage of healthcare data, leading to data being in silos (i.e. locked-in at healthcare facilities). The application of federated algorithms on these datasets could accelerate…
▽ More
Background: Federated learning methods offer the possibility of training machine learning models on privacy-sensitive data sets, which cannot be easily shared. Multiple regulations pose strict requirements on the storage and usage of healthcare data, leading to data being in silos (i.e. locked-in at healthcare facilities). The application of federated algorithms on these datasets could accelerate disease diagnostic, drug development, as well as improve patient care.
Methods: We present an extensive evaluation of the impact of different federation and differential privacy techniques when training models on the open-source MIMIC-III dataset. We analyze a set of parameters influencing a federated model performance, namely data distribution (homogeneous and heterogeneous), communication strategies (communication rounds vs. local training epochs), federation strategies (FedAvg vs. FedProx). Furthermore, we assess and compare two differential privacy (DP) techniques during model training: a stochastic gradient descent-based differential privacy algorithm (DP-SGD), and a sparse vector differential privacy technique (DP-SVT).
Results: Our experiments show that extreme data distributions across sites (imbalance either in the number of patients or the positive label ratios between sites) lead to a deterioration of model performance when trained using the FedAvg strategy. This issue is resolved when using FedProx with the use of appropriate hyperparameter tuning. Furthermore, the results show that both differential privacy techniques can reach model performances similar to those of models trained without DP, however at the expense of a large quantifiable privacy leakage.
Conclusions: We evaluate empirically the benefits of two federation strategies and propose optimal strategies for the choice of parameters when using differential privacy techniques.
△ Less
Submitted 8 February, 2023;
originally announced February 2023.
-
Translation beyond Delsarte
Authors:
Á. P. Horváth
Abstract:
We introduce general translations as solutions to Cauchy or Dirichlet problems. This point of view allows us to handle the heat-diffusion semigroup as a translation. With the given examples Kolmogorov-Riesz characterization of compact sets in certain $L^p_μ$ spaces are given. Pego-type characterizations are also derived. Finally for some examples the equivalence of the corresponding modulus of smo…
▽ More
We introduce general translations as solutions to Cauchy or Dirichlet problems. This point of view allows us to handle the heat-diffusion semigroup as a translation. With the given examples Kolmogorov-Riesz characterization of compact sets in certain $L^p_μ$ spaces are given. Pego-type characterizations are also derived. Finally for some examples the equivalence of the corresponding modulus of smoothness and K-functional is pointed out.
△ Less
Submitted 14 June, 2023; v1 submitted 1 July, 2022;
originally announced July 2022.
-
Saliency Map Based Data Augmentation
Authors:
Jalal Al-afandi,
Bálint Magyar,
András Horváth
Abstract:
Data augmentation is a commonly applied technique with two seemingly related advantages. With this method one can increase the size of the training set generating new samples and also increase the invariance of the network against the applied transformations. Unfortunately all images contain both relevant and irrelevant features for classification therefore this invariance has to be class specific…
▽ More
Data augmentation is a commonly applied technique with two seemingly related advantages. With this method one can increase the size of the training set generating new samples and also increase the invariance of the network against the applied transformations. Unfortunately all images contain both relevant and irrelevant features for classification therefore this invariance has to be class specific. In this paper we will present a new method which uses saliency maps to restrict the invariance of neural networks to certain regions, providing higher test accuracy in classification tasks.
△ Less
Submitted 29 May, 2022;
originally announced May 2022.
-
On the Feasibility and Generality of Patch-based Adversarial Attacks on Semantic Segmentation Problems
Authors:
Soma Kontar,
Andras Horvath
Abstract:
Deep neural networks were applied with success in a myriad of applications, but in safety critical use cases adversarial attacks still pose a significant threat. These attacks were demonstrated on various classification and detection tasks and are usually considered general in a sense that arbitrary network outputs can be generated by them.
In this paper we will demonstrate through simple case s…
▽ More
Deep neural networks were applied with success in a myriad of applications, but in safety critical use cases adversarial attacks still pose a significant threat. These attacks were demonstrated on various classification and detection tasks and are usually considered general in a sense that arbitrary network outputs can be generated by them.
In this paper we will demonstrate through simple case studies both in simulation and in real-life, that patch based attacks can be utilised to alter the output of segmentation networks. Through a few examples and the investigation of network complexity, we will also demonstrate that the number of possible output maps which can be generated via patch-based attacks of a given size is typically smaller than the area they effect or areas which should be attacked in case of practical applications.
We will prove that based on these results most patch-based attacks cannot be general in practice, namely they can not generate arbitrary output maps or if they could, they are spatially limited and this limit is significantly smaller than the receptive field of the patches.
△ Less
Submitted 21 May, 2022;
originally announced May 2022.
-
Observing Particle Energization above the Nyquist Frequency: An Application of the Field-Particle Correlation Technique
Authors:
Sarah A. Horvath,
Gregory G. Howes,
Andrew J. McCubbin
Abstract:
The field-particle correlation technique utilizes single-point measurements to uncover signatures of various particle energization mechanisms in turbulent space plasmas. The signature of Landau damping by electrons has been found in both simulations and observations from Earth's magnetosheath using this technique, but instrumental limitations of spacecraft sampling rates present a challenge to dis…
▽ More
The field-particle correlation technique utilizes single-point measurements to uncover signatures of various particle energization mechanisms in turbulent space plasmas. The signature of Landau damping by electrons has been found in both simulations and observations from Earth's magnetosheath using this technique, but instrumental limitations of spacecraft sampling rates present a challenge to discovering the full extent of the presence of Landau damping in the solar wind. Theory predicts that field-particle correlations can recover velocity-space energization signatures even from data that is undersampled with respect to the characteristic frequencies at which the wave damping occurs. To test this hypothesis, we perform a high-resoluation gyrokinetic simulation of space plasma turbulence, confirm that it contains signatures of electron Landau damping, and then systematically reduce the time resolution of the data to identify the point at which the signatures become impossible to recover. We find results in support of our theoretical prediction and look for a rule of thumb that can be compared with the measurement capabilities of spacecraft missions to inform the process of applying field-particle correlations to low time resolution data.
△ Less
Submitted 31 March, 2022;
originally announced April 2022.
-
Border of the Island of Inversion: Unbound states in $^{29}$Ne
Authors:
M. Holl,
S. Lindberg,
A. Heinz,
Y. Kondo,
T. Nakamura,
J. A. Tostevin,
H. Wang,
T. Nilsson,
N. L. Achouri,
H. Al Falou,
L. Atar,
T. Aumann,
H. Baba,
K. Boretzky,
C. Caesar,
D. Calvet,
H. Chae,
N. Chiga,
A. Corsi,
H. L. Crawford,
F. Delaunay,
A. Delbart,
Q. Deshayes,
P. Díaz Fernández,
Z. Dombrádi
, et al. (67 additional authors not shown)
Abstract:
The nucleus $^{29}$Ne is situated at the border of the island of inversion. Despite significant efforts, no bound low-lying intruder $f_{7/2}$-state, which would place $^{29}$Ne firmly inside the island of inversion, has yet been observed. Here, the first investigation of unbound states of $^{29}$Ne is reported. The states were populated in $^{30}\mathrm{Ne}(p,pn)$ and $^{30}\mathrm{Na}(p,2p)$ rea…
▽ More
The nucleus $^{29}$Ne is situated at the border of the island of inversion. Despite significant efforts, no bound low-lying intruder $f_{7/2}$-state, which would place $^{29}$Ne firmly inside the island of inversion, has yet been observed. Here, the first investigation of unbound states of $^{29}$Ne is reported. The states were populated in $^{30}\mathrm{Ne}(p,pn)$ and $^{30}\mathrm{Na}(p,2p)$ reactions at a beam energy of around $230$ MeV/nucleon, and analyzed in terms of their resonance properties, partial cross sections and momentum distributions. The momentum distributions are compared to calculations using the eikonal, direct reaction model, allowing $\ell$-assignments for the observed states. The lowest-lying resonance at an excitation energy of 1.48(4) MeV shows clear signs of a significant $\ell$=3-component, giving first evidence for $f_{7/2}$ single particle strength in $^{29}$Ne. The excitation energies and strengths of the observed states are compared to shell-model calculations using the sdpf-u-mix interaction
△ Less
Submitted 11 February, 2022;
originally announced February 2022.
-
Mitigating the Bias of Centered Objects in Common Datasets
Authors:
Gergely Szabo,
Andras Horvath
Abstract:
Convolutional networks are considered shift invariant, but it was demonstrated that their response may vary according to the exact location of the objects. In this paper we will demonstrate that most commonly investigated datasets have a bias, where objects are over-represented at the center of the image during training. This bias and the boundary condition of these networks can have a significant…
▽ More
Convolutional networks are considered shift invariant, but it was demonstrated that their response may vary according to the exact location of the objects. In this paper we will demonstrate that most commonly investigated datasets have a bias, where objects are over-represented at the center of the image during training. This bias and the boundary condition of these networks can have a significant effect on the performance of these architectures and their accuracy drops significantly as an object approaches the boundary. We will also demonstrate how this effect can be mitigated with data augmentation techniques.
△ Less
Submitted 4 August, 2023; v1 submitted 16 December, 2021;
originally announced December 2021.
-
Understanding How Programmers Can Use Annotations on Documentation
Authors:
Amber Horvath,
Michael Xieyang Liu,
River Hendriksen,
Connor Shannon,
Emma Paterson,
Kazi Jawad,
Andrew Macvean,
Brad A. Myers
Abstract:
Modern software development requires developers to find and effectively utilize new APIs and their documentation, but documentation has many well-known issues. Despite this, developers eventually overcome these issues but have no way of sharing what they learned. We investigate sharing this documentation-specific information through \textit{annotations}, which have advantages over developer forums…
▽ More
Modern software development requires developers to find and effectively utilize new APIs and their documentation, but documentation has many well-known issues. Despite this, developers eventually overcome these issues but have no way of sharing what they learned. We investigate sharing this documentation-specific information through \textit{annotations}, which have advantages over developer forums as the information is contextualized, not disruptive, and is short, thus easy to author. Developers can also author annotations to support their own comprehension. In order to support the documentation usage behaviors we found, we built the Adamite annotation tool, which supports features such as multi-anchoring, annotation types, and pinning. In our user study, we found that developers are able to create annotations that are useful to themselves and are able to utilize annotations created by other developers when learning a new API, with readers of the annotations completing 67% more of the task, on average, than the baseline.
△ Less
Submitted 11 January, 2022; v1 submitted 16 November, 2021;
originally announced November 2021.
-
A two-vertex theorem for normal tilings
Authors:
Gábor Domokos,
Ákos G. Horváth,
Krisztina Regős
Abstract:
We regard a smooth, $d=2$-dimensional manifold $\mathcal{M}$ and its normal tiling $M$, the cells of which may have non-smooth or smooth vertices (at the latter, two edges meet at 180 degrees.) We denote the average number (per cell) of non-smooth vertices by $\bar v^{\star}$ and we prove that if $M$ is periodic then $v^{\star} \geq 2$ and we show the same result for the monohedral case by an enti…
▽ More
We regard a smooth, $d=2$-dimensional manifold $\mathcal{M}$ and its normal tiling $M$, the cells of which may have non-smooth or smooth vertices (at the latter, two edges meet at 180 degrees.) We denote the average number (per cell) of non-smooth vertices by $\bar v^{\star}$ and we prove that if $M$ is periodic then $v^{\star} \geq 2$ and we show the same result for the monohedral case by an entirely different argument. Our theory also makes a closely related prediction for non-periodic tilings. In 3 dimensions we show a monohedral construction with $\bar v^{\star}=0$.
△ Less
Submitted 5 January, 2022; v1 submitted 5 October, 2021;
originally announced October 2021.
-
Compactness criteria via Laguerre and Hankel transformations
Authors:
Á. P. Horváth
Abstract:
The aim of this paper is to prove Kolmogorov-Riesz type theorems via Bessel and Laguerre translations, and Pego-type theorems by the corresponding transformations.
The aim of this paper is to prove Kolmogorov-Riesz type theorems via Bessel and Laguerre translations, and Pego-type theorems by the corresponding transformations.
△ Less
Submitted 15 March, 2021;
originally announced March 2021.
-
On the coordinates of minimal vectors in a Minkowski-reduced basis
Authors:
Ákos G. Horváth
Abstract:
Finding the shortest vectors in a lattice is an NP-hard problem, so low-dimensional results also play an essential role in lattice reduction theory. Using Ryskov's result for the admissible centerings and Tammela's result for determining the Minkowski-reduced form, we prove that the absolute values of the coordinates of a minimal vector on a six-dimensional Minkowski-reduced basis are less than or…
▽ More
Finding the shortest vectors in a lattice is an NP-hard problem, so low-dimensional results also play an essential role in lattice reduction theory. Using Ryskov's result for the admissible centerings and Tammela's result for determining the Minkowski-reduced form, we prove that the absolute values of the coordinates of a minimal vector on a six-dimensional Minkowski-reduced basis are less than or equal to three. To sharpen P. Tammela's work, we combine some lattice geometry arguments with the aforementioned theoretical results.
△ Less
Submitted 10 October, 2024; v1 submitted 9 February, 2021;
originally announced February 2021.
-
On the convex hull and homothetic convex hull functions of a convex body
Authors:
Ákos G. Horváth,
Zsolt Lángi
Abstract:
The aim of this note is to investigate the properties of the convex hull and the homothetic convex hull functions of a convex body $K$ in Euclidean $n$-space, defined as the volume of the union of $K$ and one of its translates, and the volume of $K$ and a translate of a homothetic copy of $K$, respectively, as functions of the translation vector. In particular, we prove that the convex hull functi…
▽ More
The aim of this note is to investigate the properties of the convex hull and the homothetic convex hull functions of a convex body $K$ in Euclidean $n$-space, defined as the volume of the union of $K$ and one of its translates, and the volume of $K$ and a translate of a homothetic copy of $K$, respectively, as functions of the translation vector. In particular, we prove that the convex hull function of the body $K$ does not determine $K$. Furthermore, we prove the equivalence of the polar projection body problem raised by Petty, and a conjecture of G.Horváth and Lángi about translative constant volume property of convex bodies. We give a short proof of some theorems of Jerónimo-Castro about the homothetic convex hull function, and prove a homothetic variant of the translative constant volume property conjecture for $3$-dimensional convex polyhedra. We also apply our results to describe the properties of the illumination bodies of convex bodies.
△ Less
Submitted 23 September, 2021; v1 submitted 16 December, 2020;
originally announced December 2020.
-
Diameter, width and thickness in the hyperbolic plane
Authors:
Ákos G. Horváth
Abstract:
This paper contains a new concept to measure the width and thickness of a convex body in the hyperbolic plane. We compare the known concepts with the new one and prove some results on bodies of constant width, constant diameter and given thickness.
This paper contains a new concept to measure the width and thickness of a convex body in the hyperbolic plane. We compare the known concepts with the new one and prove some results on bodies of constant width, constant diameter and given thickness.
△ Less
Submitted 30 November, 2020;
originally announced November 2020.
-
Note on the Equilibrium Measures of Julia sets of Exceptional Jacobi Polynomials
Authors:
Á. P. Horváth
Abstract:
We prove that similarly to the standard case, the equilibrium measure of Julia sets of exceptional Jacobi polynomials tends to the equilibrium measure of the interval of orthogonality in weak-star sense.
We prove that similarly to the standard case, the equilibrium measure of Julia sets of exceptional Jacobi polynomials tends to the equilibrium measure of the interval of orthogonality in weak-star sense.
△ Less
Submitted 16 November, 2020;
originally announced November 2020.
-
Receptive Field Size Optimization with Continuous Time Pooling
Authors:
Dóra Babicz,
Soma Kontár,
Márk Pető,
András Fülöp,
Gergely Szabó,
András Horváth
Abstract:
The pooling operation is a cornerstone element of convolutional neural networks. These elements generate receptive fields for neurons, in which local perturbations should have minimal effect on the output activations, increasing robustness and invariance of the network. In this paper we will present an altered version of the most commonly applied method, maximum pooling, where pooling in theory is…
▽ More
The pooling operation is a cornerstone element of convolutional neural networks. These elements generate receptive fields for neurons, in which local perturbations should have minimal effect on the output activations, increasing robustness and invariance of the network. In this paper we will present an altered version of the most commonly applied method, maximum pooling, where pooling in theory is substituted by a continuous time differential equation, which generates a location sensitive pooling operation, more similar to biological receptive fields. We will present how this continuous method can be approximated numerically using discrete operations which fit ideally on a GPU. In our approach the kernel size is substituted by diffusion strength which is a continuous valued parameter, this way it can be optimized by gradient descent algorithms. We will evaluate the effect of continuous pooling on accuracy and computational need using commonly applied network architectures and datasets.
△ Less
Submitted 6 November, 2020; v1 submitted 2 November, 2020;
originally announced November 2020.
-
Filtered Batch Normalization
Authors:
Andras Horvath,
Jalal Al-afandi
Abstract:
It is a common assumption that the activation of different layers in neural networks follow Gaussian distribution. This distribution can be transformed using normalization techniques, such as batch-normalization, increasing convergence speed and improving accuracy. In this paper we would like to demonstrate, that activations do not necessarily follow Gaussian distribution in all layers. Neurons in…
▽ More
It is a common assumption that the activation of different layers in neural networks follow Gaussian distribution. This distribution can be transformed using normalization techniques, such as batch-normalization, increasing convergence speed and improving accuracy. In this paper we would like to demonstrate, that activations do not necessarily follow Gaussian distribution in all layers. Neurons in deeper layers are more selective and specific which can result extremely large, out-of-distribution activations.
We will demonstrate that one can create more consistent mean and variance values for batch normalization during training by filtering out these activations which can further improve convergence speed and yield higher validation accuracy.
△ Less
Submitted 16 October, 2020;
originally announced October 2020.
-
3D Segmentation Networks for Excessive Numbers of Classes: Distinct Bone Segmentation in Upper Bodies
Authors:
Eva Schnider,
Antal Horváth,
Georg Rauter,
Azhar Zam,
Magdalena Müller-Gerbl,
Philippe C. Cattin
Abstract:
Segmentation of distinct bones plays a crucial role in diagnosis, planning, navigation, and the assessment of bone metastasis. It supplies semantic knowledge to visualisation tools for the planning of surgical interventions and the education of health professionals. Fully supervised segmentation of 3D data using Deep Learning methods has been extensively studied for many tasks but is usually restr…
▽ More
Segmentation of distinct bones plays a crucial role in diagnosis, planning, navigation, and the assessment of bone metastasis. It supplies semantic knowledge to visualisation tools for the planning of surgical interventions and the education of health professionals. Fully supervised segmentation of 3D data using Deep Learning methods has been extensively studied for many tasks but is usually restricted to distinguishing only a handful of classes. With 125 distinct bones, our case includes many more labels than typical 3D segmentation tasks. For this reason, the direct adaptation of most established methods is not possible. This paper discusses the intricacies of training a 3D segmentation network in a many-label setting and shows necessary modifications in network architecture, loss function, and data augmentation. As a result, we demonstrate the robustness of our method by automatically segmenting over one hundred distinct bones simultaneously in an end-to-end learnt fashion from a CT-scan.
△ Less
Submitted 14 October, 2020;
originally announced October 2020.
-
Electron Landau Damping of Kinetic Alfvén Waves in Simulated Magnetosheath Turbulence
Authors:
Sarah A. Horvath,
Gregory G. Howes,
Andrew J. McCubbin
Abstract:
Turbulence is thought to play a role in the heating of the solar wind plasma, though many questions remain to be solved regarding the exact nature of the mechanisms driving this process in the heliosphere. In particular, the physics of the collisionless interactions between particles and turbulent electromagnetic fields in the kinetic dissipation range of the turbulent cascade remains incompletely…
▽ More
Turbulence is thought to play a role in the heating of the solar wind plasma, though many questions remain to be solved regarding the exact nature of the mechanisms driving this process in the heliosphere. In particular, the physics of the collisionless interactions between particles and turbulent electromagnetic fields in the kinetic dissipation range of the turbulent cascade remains incompletely understood. A recent analysis of an interval of Magnetosphere Multiscale (MMS) observations has used the field-particle correlation technique to demonstrate that electron Landau damping is involved in the dissipation of turbulence in the Earth's magnetosheath. Motivated by this discovery, we perform a high-resolution gyrokinetic numerical simulation of the turbulence in the MMS interval to investigate the role of electron Landau damping in the dissipation of turbulent energy. We employ the field-particle correlation technique on our simulation data, compare our results to the known velocity-space signatures of Landau damping outside the dissipation range, and evaluate the net electron energization. We find qualitative agreement between the numerical and observational results for some key aspects of the energization and speculate on the nature of disagreements in light of experimental factors, such as differences in resolution, and of developing insights into the nature of field-particle interactions in the presence of dispersive kinetic Alfvén waves.
△ Less
Submitted 10 September, 2020;
originally announced September 2020.
-
Sorted Pooling in Convolutional Networks for One-shot Learning
Authors:
András Horváth
Abstract:
We present generalized versions of the commonly used maximum pooling operation: $k$th maximum and sorted pooling operations which selects the $k$th largest response in each pooling region, selecting locally consistent features of the input images. This method is able to increase the generalization power of a network and can be used to decrease training time and error rate of networks and it can si…
▽ More
We present generalized versions of the commonly used maximum pooling operation: $k$th maximum and sorted pooling operations which selects the $k$th largest response in each pooling region, selecting locally consistent features of the input images. This method is able to increase the generalization power of a network and can be used to decrease training time and error rate of networks and it can significantly improve accuracy in case of training scenarios where the amount of available data is limited, like one-shot learning scenarios
△ Less
Submitted 20 July, 2020;
originally announced July 2020.
-
Discrete diffusion semigroups associated with Dunkl-Jacobi and exceptional Jacobi polynomials
Authors:
Á. P. Horváth
Abstract:
Some weighted inequalities for the maximal operator with respect to the discrete diffusion semigroups associated with exceptional Jacobi and Dunkl-Jacobi polynomials are given. This setup allows to extend the corresponding results obtained for discrete heat semigroup recently to richer class of differential-difference operators.
Some weighted inequalities for the maximal operator with respect to the discrete diffusion semigroups associated with exceptional Jacobi and Dunkl-Jacobi polynomials are given. This setup allows to extend the corresponding results obtained for discrete heat semigroup recently to richer class of differential-difference operators.
△ Less
Submitted 11 July, 2020;
originally announced July 2020.
-
Deposition distribution of the new coronavirus (SARS-CoV-2) in the human airways upon exposure to cough-generated aerosol
Authors:
Balázs G. Madas,
Péter Füri,
Árpád Farkas,
Attila Nagy,
Aladár Czitrovszky,
Imre Balásházy,
Gusztáv G. Schay,
Alpár Horváth
Abstract:
The new coronavirus disease 2019 (COVID-19) has been emerged as a rapidly spreading pandemic. The disease is thought to spread mainly from person-to-person through respiratory droplets produced when an infected person coughs, sneezes, or talks. The pathogen of COVID-19 is the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). It infects the cells binding to the angiotensin-converting en…
▽ More
The new coronavirus disease 2019 (COVID-19) has been emerged as a rapidly spreading pandemic. The disease is thought to spread mainly from person-to-person through respiratory droplets produced when an infected person coughs, sneezes, or talks. The pathogen of COVID-19 is the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). It infects the cells binding to the angiotensin-converting enzyme 2 receptor (ACE2) which is expressed by cells throughout the airways as targets for cellular entry. Although the majority of persons infected with SARS-CoV-2 experience symptoms of mild upper respiratory tract infection, in some people infections of the peripheral airways result in severe, potentially fatal pneumonia. However, the induction of COVID-19 pneumonia requires that SARS-CoV-2 reaches the peripheral airways. While huge efforts have been made to understand the spread of the disease as well as the pathogenesis following cellular entry, much less attention is paid how SARS-CoV-2 from the environment reach the receptors of the target cells. The aim of the present study is to characterize the deposition distribution of SARS-CoV-2 in the airways upon exposure to cough-generated aerosol. For this purpose, the Stochastic Lung Deposition Model has been applied. Aerosol size distribution and breathing parameters were taken from the literature supposing normal breathing through the nose. We found that the probability of direct infection of the peripheral airways due to inhalation of aerosol generated by a bystander cough is very low. As the number of pathogens deposited in the extrathoracic airways is ~10 times higher than in the peripheral airways, we concluded that in most cases COVID-19 pneumonia must be preceded by SARS-CoV-2 infection of the upper airways. Our results suggest that without the enhancement of viral load in the upper airways, COVID-19 would be much less dangerous...
△ Less
Submitted 12 May, 2020;
originally announced May 2020.
-
Extending the Southern Shore of the Island of Inversion to $^{28}$F
Authors:
A. Revel,
O. Sorlin,
F. M. Marques,
Y. Kondo,
J. Kahlbow,
T. Nakamura,
N. A. Orr,
F. Nowacki,
J. A. Tostevin,
C. X. Yuan,
N. L. Achouri,
H. Al Falou,
L. Atar,
T. Aumann,
H. Baba,
K. Boretzky,
C. Caesar,
D. Calvet,
H. Chae,
N. Chiga,
A. Corsi,
H. L. Crawford,
F. Delaunay,
A. Delbart,
Q. Deshayes
, et al. (67 additional authors not shown)
Abstract:
Detailed spectroscopy of the neutron-unbound nucleus $^{28}$F has been performed for the first time following proton/neutron removal from $^{29}$Ne/$^{29}$F beams at energies around 230 MeV/nucleon. The invariant-mass spectra were reconstructed for both the $^{27}$F$^{(*)}+n$ and $^{26}$F$^{(*)}+2n$ coincidences and revealed a series of well-defined resonances. A near-threshold state was observed…
▽ More
Detailed spectroscopy of the neutron-unbound nucleus $^{28}$F has been performed for the first time following proton/neutron removal from $^{29}$Ne/$^{29}$F beams at energies around 230 MeV/nucleon. The invariant-mass spectra were reconstructed for both the $^{27}$F$^{(*)}+n$ and $^{26}$F$^{(*)}+2n$ coincidences and revealed a series of well-defined resonances. A near-threshold state was observed in both reactions and is identified as the $^{28}$F ground state, with $S_n(^{28}$F$)=-199(6)$ keV, while analysis of the $2n$ decay channel allowed a considerably improved $S_n(^{27}$F$)=1620(60)$ keV to be deduced. Comparison with shell-model predictions and eikonal-model reaction calculations have allowed spin-parity assignments to be proposed for some of the lower-lying levels of $^{28}$F. Importantly, in the case of the ground state, the reconstructed $^{27}$F$+n$ momentum distribution following neutron removal from $^{29}$F indicates that it arises mainly from the $1p_{3/2}$ neutron intruder configuration. This demonstrates that the island of inversion around $N=20$ includes $^{28}$F, and most probably $^{29}$F, and suggests that $^{28}$O is not doubly magic.
△ Less
Submitted 2 April, 2020;
originally announced April 2020.
-
Multiplication operator and exceptional Jacobi polynomials
Authors:
Á. P. Horváth
Abstract:
Below the normalized weighted reciprocal of the Christoffel function with respect to exceptional Jacobi polynomials is investigated. It is proved that it tends to the equilibrium measure of the interval of orthogonality in weak-star sense. The main tool of this study is the multiplication operator and examination of behavior of zeros of the corresponding average characteristic polynomial. Finally,…
▽ More
Below the normalized weighted reciprocal of the Christoffel function with respect to exceptional Jacobi polynomials is investigated. It is proved that it tends to the equilibrium measure of the interval of orthogonality in weak-star sense. The main tool of this study is the multiplication operator and examination of behavior of zeros of the corresponding average characteristic polynomial. Finally, as an application of multiplication operator, location of zeros of certain self-inversive polynomials are examined.
△ Less
Submitted 16 November, 2020; v1 submitted 26 March, 2020;
originally announced March 2020.
