arXiv subjects

Thomas Weber

Publications and source records attributed to Thomas Weber.

At least 55 records · Page 3Linked to original sources

arXiv2022

Permittivity-asymmetric quasi-bound states in the continuum

Broken symmetries lie at the heart of nontrivial physical phenomena. Breaking the in-plane geometrical symmetry of optical systems allows to access a set of electromagnetic states termed symmetry-protected quasi-bound states in the continuum (qBICs). Here we demonstrate, theoretically, numerically and experimentally, that such optical states can also be accessed in metasurfaces by breaking the in-plane symmetry in the permittivity of the comprising materials, showing a remarkable equivalence to their geometrically-asymmetric counterparts. However, while the physical size of atoms imposes a limit on the lowest achievable geometrical asymmetry, weak permittivity modulations due to carrier doping and electro-optical Pockels and Kerr effects, usually considered insignificant, open up the possibility of infinitesimal permittivity asymmetries for on-demand, and dynamically tuneable optical resonances of extremely high quality factors. We probe the excitation of permittivity-asymmetric qBICs (${\varepsilon}$-qBICs) using a prototype Si/TiO$_{2}$ metasurface, in which the asymmetry in the unit cell is provided by the refractive index contrast of the dissimilar materials, surpassing any unwanted asymmetries from nanofabrication defects or angular deviations of light from normal incidence. ${\varepsilon}$-qBICs can also be excited in 1D gratings, where quality-factor enhancement and tailored interference phenomena via the interplay of geometrical and permittivity asymmetries are numerically demonstrated. The emergence of ${\varepsilon}$-qBICs in systems with broken symmetries in their permittivity may enable to test time-energy uncertainties in quantum mechanics, and lead to a whole new class of low-footprint optical and optoelectronic devices, from arbitrarily narrow filters and topological sources, biosensing and ultrastrong light-matter interaction platforms, to tuneable optical switches.

physics.optics↗

arXiv2022

High-Q nanophotonics over the full visible spectrum enabled by hexagonal boron nitride metasurfaces

All-dielectric optical metasurfaces with high quality (Q) factors have so far been hampered by the lack of simultaneously lossless and high refractive index (RI) materials over the full visible spectrum. To achieve broad spectral coverage, the use of low-index materials is, in fact, unavoidable due to the inverse correlation between the band-gap energy (and therefore the optical losses) and the RI. However, for Mie resonant photonics, smaller RIs are associated with reduced Q factors and mode volume confinement. In this work, we leverage symmetry-broken bound states in the continuum (BICs) to efficiently suppress radiation losses from the low-index (n~2) van der Waals material hexagonal boron nitride (hBN), realizing metasurfaces with high-Q resonances over the complete visible spectrum. In particular, we analyze the rational use of low and high RI materials as resonator components and harness our insights to experimentally demonstrate sharp BIC resonances with Q factors above 300, spanning wavelengths between 400 nm and 1000 nm from a single hBN flake. Moreover, we utilize the enhanced electric near-fields to demonstrate second harmonic generation (SHG) with enhancement factors above 102. Our results provide a theoretical and experimental framework for the implementation of low RI materials as photonic media for metaoptics.

physics.optics↗

arXiv2022

Strong light-matter interaction with self-hybridized bound states in the continuum in monolithic van der Waals metasurfaces

Photonic bound states in the continuum (BICs) are a standout nanophotonic platform for strong light-matter coupling with transition metal dichalcogenides (TMDCs), but have so far mostly been employed as all-dielectric metasurfaces with adjacent TMDC layers, incurring limitations related to strain, mode overlap, and material integration. In this work, we experimentally demonstrate for the first time asymmetry-dependent BIC resonances in 2D arrays of monolithic metasurfaces composed solely of the nanostructured bulk TMDC WS$_2$ with BIC modes exhibiting sharp and tailored linewidths, ideal for selectively enhancing light-matter interactions. Geometrical variation enables the tuning of the BIC resonances across the exciton resonance in bulk WS$_2$, revealing the strong-coupling regime with an anti-crossing pattern and a Rabi splitting of 116 meV. The precise control over the radiative loss channel provided by the BIC concept is harnessed to tailor the Rabi splitting via a geometrical asymmetry parameter of the metasurface. Crucially, the coupling strength itself can be controlled and is shown to be independent of material-intrinsic losses. Our BIC-driven monolithic metasurface platform can readily incorporate other TMDCs or excitonic materials to deliver previously unavailable fundamental insights and practical device concepts for polaritonic applications.

physics.optics↗

arXiv2022

Plasmonic Bound States in the Continuum to Tailor Light-Matter Coupling

Plasmon resonances play a pivotal role in enhancing light-matter interactions in nanophotonics, but their low-quality factors have hindered applications demanding high spectral selectivity. Even though symmetry-protected bound states in the continuum with high-quality factors have been realized in dielectric metasurfaces, impinging light is not efficiently coupled to the resonant metasurfaces and is lost in the form of reflection due to low intrinsic losses. Here, we demonstrate a novel design and 3D laser nanoprinting of plasmonic nanofin metasurfaces, which support symmetry-protected bound states in the continuum up to 4th order. By breaking the nanofins out-of-plane symmetry in parameter space, we achieve high-quality factor (up to 180) modes under normal incidence. We reveal that the out-of-plane symmetry breaking can be fine-tuned by the triangle angle of the 3D nanofin meta-atoms, opening a pathway to precisely control the ratio of radiative to intrinsic losses. This enables access to the under-, critical-, and over-coupled regimes, which we exploit for pixelated molecular sensing. Depending on the coupling regime we observe negative, no, or positive modulation induced by the analyte, unveiling the undeniable importance of tailoring light-matter interaction. Our demonstration provides a novel metasurface platform for enhanced light-matter interaction with a wide range of applications in optical sensing, energy conversion, nonlinear photonics, surface-enhanced spectroscopy, and quantum optics.

physics.optics↗

arXiv2022

Twisted geometry for submanifolds of $\mathbb{R}^n$

This is a friendly introduction to our recent general procedure for constructing noncommutative deformations of an embedded submanifold $M$ of $\mathbb{R}^n$ determined by a set of smooth equations $f^a(x)=0$. We use the framework of Drinfel'd twist deformation of differential geometry pioneered in [Aschieri et al., Class. Quantum Gravity 23 (2006), 1883]; the commutative pointwise product is replaced by a (generally noncommutative) $\star$-product induced by a Drinfel'd twist.

math-ph↗

arXiv2021

Ethically aligned Deep Learning: Unbiased Facial Aesthetic Prediction

Facial beauty prediction (FBP) aims to develop a machine that automatically makes facial attractiveness assessment. In the past those results were highly correlated with human ratings, therefore also with their bias in annotating. As artificial intelligence can have racist and discriminatory tendencies, the cause of skews in the data must be identified. Development of training data and AI algorithms that are robust against biased information is a new challenge for scientists. As aesthetic judgement usually is biased, we want to take it one step further and propose an Unbiased Convolutional Neural Network for FBP. While it is possible to create network models that can rate attractiveness of faces on a high level, from an ethical point of view, it is equally important to make sure the model is unbiased. In this work, we introduce AestheticNet, a state-of-the-art attractiveness prediction network, which significantly outperforms competitors with a Pearson Correlation of 0.9601. Additionally, we propose a new approach for generating a bias-free CNN to improve fairness in machine learning.

cs.CV↗

arXiv2021

Twisted submanifolds of R^n

We propose a general procedure to construct noncommutative deformations of an embedded submanifold $M$ of $\mathbb{R}^n$ determined by a set of smooth equations $f^a(x)=0$. We use the framework of Drinfel'd twist deformation of differential geometry of [Aschieri et al., Class. Quantum Gravity 23 (2006), 1883]; the commutative pointwise product is replaced by a (generally noncommutative) $\star$-product determined by a Drinfel'd twist. The twists we employ are based on the Lie algebra $Ξ_t$ of vector fields that are tangent to all the submanifolds that are level sets of the $f^a$; the twisted Cartan calculus is automatically equivariant under twisted tangent infinitesimal diffeomorphisms. We can consistently project a connection from the twisted $\mathbb{R}^n$ to the twisted $M$ if the twist is based on a suitable Lie subalgebra $\mathfrak{e}\subsetΞ_t$. If we endow $\mathbb{R}^n$ with a metric then twisting and projecting to the normal and tangent vector fields commute, and we can project the Levi-Civita connection consistently to the twisted $M$, provided the twist is based on the Lie subalgebra $\mathfrak{k}\subset\mathfrak{e}$ of the Killing vector fields of the metric; a twisted Gauss theorem follows, in particular. Twisted algebraic manifolds can be characterized in terms of generators and polynomial relations. We present in some detail twisted cylinders embedded in twisted Euclidean $\mathbb{R}^3$ and twisted hyperboloids embedded in twisted Minkowski $\mathbb{R}^3$ [these are twisted (anti-)de Sitter spaces $dS_2,AdS_2$].

math-ph↗

arXiv2021

Metapath- and Entity-aware Graph Neural Network for Recommendation

In graph neural networks (GNNs), message passing iteratively aggregates nodes' information from their direct neighbors while neglecting the sequential nature of multi-hop node connections. Such sequential node connections e.g., metapaths, capture critical insights for downstream tasks. Concretely, in recommender systems (RSs), disregarding these insights leads to inadequate distillation of collaborative signals. In this paper, we employ collaborative subgraphs (CSGs) and metapaths to form metapath-aware subgraphs, which explicitly capture sequential semantics in graph structures. We propose meta\textbf{P}ath and \textbf{E}ntity-\textbf{A}ware \textbf{G}raph \textbf{N}eural \textbf{N}etwork (PEAGNN), which trains multilayer GNNs to perform metapath-aware information aggregation on such subgraphs. This aggregated information from different metapaths is then fused using attention mechanism. Finally, PEAGNN gives us the representations for node and subgraph, which can be used to train MLP for predicting score for target user-item pairs. To leverage the local structure of CSGs, we present entity-awareness that acts as a contrastive regularizer on node embedding. Moreover, PEAGNN can be combined with prominent layers such as GAT, GCN and GraphSage. Our empirical evaluation shows that our proposed technique outperforms competitive baselines on several datasets for recommendation tasks. Further analysis demonstrates that PEAGNN also learns meaningful metapath combinations from a given set of metapaths.

cs.LG↗

arXiv2020

Discovering key topics from short, real-world medical inquiries via natural language processing and unsupervised learning

Millions of unsolicited medical inquiries are received by pharmaceutical companies every year. It has been hypothesized that these inquiries represent a treasure trove of information, potentially giving insight into matters regarding medicinal products and the associated medical treatments. However, due to the large volume and specialized nature of the inquiries, it is difficult to perform timely, recurrent, and comprehensive analyses. Here, we propose a machine learning approach based on natural language processing and unsupervised learning to automatically discover key topics in real-world medical inquiries from customers. This approach does not require ontologies nor annotations. The discovered topics are meaningful and medically relevant, as judged by medical information specialists, thus demonstrating that unsolicited medical inquiries are a source of valuable customer insights. Our work paves the way for the machine-learning-driven analysis of medical inquiries in the pharmaceutical industry, which ultimately aims at improving patient care.

cs.LG↗

arXiv2020

Encoder-decoder semantic segmentation models for electroluminescence images of thin-film photovoltaic modules

We consider a series of image segmentation methods based on the deep neural networks in order to perform semantic segmentation of electroluminescence (EL) images of thin-film modules. We utilize the encoder-decoder deep neural network architecture. The framework is general such that it can easily be extended to other types of images (e.g. thermography) or solar cell technologies (e.g. crystalline silicon modules). The networks are trained and tested on a sample of images from a database with 6000 EL images of Copper Indium Gallium Diselenide (CIGS) thin film modules. We selected two types of features to extract, shunts and so called "droplets". The latter feature is often observed in the set of images. Several models are tested using various combinations of encoder-decoder layers, and a procedure is proposed to select the best model. We show exemplary results with the best selected model. Furthermore, we applied the best model to the full set of 6000 images and demonstrate that the automated segmentation of EL images can reveal many subtle features which cannot be inferred from studying a small sample of images. We believe these features can contribute to process optimization and quality control.

eess.IV↗

arXiv2020

Twisted Quadrics and Algebraic Submanifolds in $\mathbb{R}^n$

We propose a general procedure to construct noncommutative deformations of an algebraic submanifold $M$ of $\mathbb{R}^n$, specializing the procedure [G. Fiore, T. Weber, Twisted submanifolds of $\mathbb{R}^n$, arXiv:2003.03854] valid for smooth submanifolds. We use the framework of twisted differential geometry of Aschieri et al. (Class. Quantum Grav. 23, 1883-1911, 2006), whereby the commutative pointwise product is replaced by the $\star$-product determined by a Drinfel'd twist. We actually simultaneously construct noncommutative deformations of all the algebraic submanifolds $M_c$ that are level sets of the $f^a(x)$, where $f^a(x)=0$ are the polynomial equations solved by the points of $M$, employing twists based on the Lie algebra $Ξ_t$ of vector fields that are tangent to all the $M_c$. The twisted Cartan calculus is automatically equivariant under twisted $Ξ_t$. If we endow $\mathbb{R}^n$ with a metric, then twisting and projecting to normal or tangent components commute, projecting the Levi-Civita connection to the twisted $M$ is consistent, and in particular a twisted Gauss theorem holds, provided the twist is based on Killing vector fields. Twisted algebraic quadrics can be characterized in terms of generators and $\star$-polynomial relations. We explicitly work out deformations based on abelian or Jordanian twists of all quadrics in $\mathbb{R}^3$ except ellipsoids, in particular twisted cylinders embedded in twisted Euclidean $\mathbb{R}^3$ and twisted hyperboloids embedded in twisted Minkowski $\mathbb{R}^3$ [the latter are twisted (anti-)de Sitter spaces $dS_2$, $AdS_2$].

math-ph↗

arXiv2020

Braided Commutative Geometry and Drinfel'd Twist Deformations

In this thesis we give obstructions for Drinfel'd twist deformation quantization on several classes of symplectic manifolds. Motivated from this quantization procedure, we further construct a noncommutative Cartan calculus on any braided commutative algebra, as well as an equivariant Levi-Civita covariant derivative for any non-degenerate equivariant metric. This generalizes and unifies the Cartan calculus on a smooth manifold and the Cartan calculus on twist star product algebras. We prove that the Drinfel'd functor leads to equivalence classes in braided commutative geometry and commutes with submanifold algebra projection.

math.QA↗

arXiv2020

Braided Cartan Calculi and Submanifold Algebras

We construct a noncommutative Cartan calculus on any braided commutative algebra and study its applications in noncommutative geometry. The braided Lie derivative, insertion and de Rham differential are introduced and related via graded braided commutators, also incorporating the braided Schouten-Nijenhuis bracket. The resulting braided Cartan calculus generalizes the Cartan calculus on smooth manifolds and the twisted Cartan calculus. While it is a necessity of derivation based Cartan calculi on noncommutative algebras to employ central bimodules our approach allows to consider bimodules over the full underlying algebra. Furthermore, equivariant covariant derivatives and metrics on braided commutative algebras are discussed. In particular, we prove the existence and uniqueness of an equivariant Levi-Civita covariant derivative for any fixed non-degenerate equivariant metric. Operating in a symmetric braided monoidal category we argue that Drinfel'd twist deformation corresponds to gauge equivalences of braided Cartan calculi. The notions of equivariant covariant derivative and metric are compatible with the Drinfel'd functor as well. Moreover, we project braided Cartan calculi to submanifold algebras and prove that this process commutes with twist deformation.

math.QA↗

arXiv2019

Interactive Image Restoration

Machine learning and many of its applications are considered hard to approach due to their complexity and lack of transparency. One mission of human-centric machine learning is to improve algorithm transparency and user satisfaction while ensuring an acceptable task accuracy. In this work, we present an interactive image restoration framework, which exploits both image prior and human painting knowledge in an iterative manner such that they can boost on each other. Additionally, in this system users can repeatedly get feedback of their interactions from the restoration progress. This informs the users about their impact on the restoration results, which leads to better sense of control, which can lead to greater trust and approachability. The positive results of both objective and subjective evaluation indicate that, our interactive approach positively contributes to the approachability of restoration algorithms in terms of algorithm performance and user experience.

cs.HC↗

arXiv2019

Stochastic and Coarse-Grained Two-Dimensional Modeling of Directional Particle Movement

We study the evolution of interacting groups of agents in two-dimensional geometries. We introduce a microscopic stochastic model that includes floor fields modeling the global flow of individual groups as well as local interaction rules. From this microscopic model we derive an analytically-tractable system of conservation laws that governs the evolution of the macroscopic densities. Numerical simulations show good agreement between the system of conservation laws and the microscopic model, though the latter is slightly more diffusive. We conclude by deriving second-order corrections to the system of conservation laws.

physics.soc-ph↗

arXiv2018

Phase-Sensitive Region-of-Interest Computed Tomography

X-Ray Phase-Contrast Imaging (PCI) yields absorption, differential phase, and dark-field images. Computed Tomography (CT) of grating-based PCI can in principle provide high-resolution soft-tissue contrast. Recently, grating-based PCI took several hurdles towards clinical implementation by addressing, for example, acquisition speed, high X-ray energies, and system vibrations. However, a critical impediment in all grating-based systems lies in limits that constrain the grating diameter to few centimeters. In this work, we propose a system and a reconstruction algorithm to circumvent this constraint in a clinically compatible way. We propose to perform a phase-sensitive Region-of-Interest (ROI) CT within a full-field absorption CT. The biggest advantage of this approach is that it allows to correct for phase truncation artifacts, and to obtain quantitative phase values. Our method is robust, and shows high-quality results on simulated data and on a biological mouse sample. This work is a proof of concept showing the potential to use PCI in CT on large specimen, such as humans, in clinical applications.

physics.med-ph↗

arXiv2018

Methodology to analyze the accuracy of 3D objects reconstructed with collaborative robot based monocular LSD-SLAM

SLAM systems are mainly applied for robot navigation while research on feasibility for motion planning with SLAM for tasks like bin-picking, is scarce. Accurate 3D reconstruction of objects and environments is important for planning motion and computing optimal gripper pose to grasp objects. In this work, we propose the methods to analyze the accuracy of a 3D environment reconstructed using a LSD-SLAM system with a monocular camera mounted onto the gripper of a collaborative robot. We discuss and propose a solution to the pose space conversion problem. Finally, we present several criteria to analyze the 3D reconstruction accuracy. These could be used as guidelines to improve the accuracy of 3D reconstructions with monocular LSD-SLAM and other SLAM based solutions.

cs.CV↗

arXiv2017

Correlated local dipoles in PbTe

We present a combined single-crystal x-ray diffuse scattering and ab-initio molecular dynamics study of lead telluride, PbTe. Well-known for its thermoelectric and narrow-gap semiconducting properties, PbTe recently achieved further notoriety following the report of an unusual off-centering of the lead atoms, accompanied by a local symmetry breaking, on heating. This observation, which was named emphanisis, ignited considerable controversy regarding the details of the underlying local structure and the appropriate interpretation of the total scattering experiments. In this study, we demonstrate close agreement between our diffuse scattering measurements and our calculations, which allows us to analyze features such as higher-order correlations that are accessible in the simulations but not experimentally. This allowed us to discover an unusual correlated local dipole formation extending over several unit cells with an associated local reduction of the cubic symmetry in both our x-ray diffuse scattering measurements and our molecular dynamics simulations. Importantly, when averaged spatially or temporally, the most probable positions for the ions are at the centers of their coordination polyhedra. Our results therefore clarify the nature of the local symmetry breaking, and reveal the source of the earlier controversy regarding the existence or absence of off-centering. Finally, we provide an interpretation of the behavior in terms of coupled soft optical and acoustic modes, which is linked also to the high thermoelectric performance of PbTe.

cond-mat.mtrl-sci↗