Search

In this work, we study the training and generalization performance of two-layer neural networks (NNs) after one gradient descent step under structured data modeled by Gaussian mixtures. While previous research has extensively analyzed this model under isotropic data assumption, such simplifications overlook the complexities inherent in real-world datasets. Our work addresses this limitation by analyzing two-layer NNs under Gaussian mixture data assumption in the asymptotically proportional limit, where the input dimension, number of hidden neurons, and sample size grow with finite ratios. We characterize the training and generalization errors by leveraging recent advancements in Gaussian universality. Specifically, we prove that a high-order polynomial model performs equivalent to the nonlinear neural networks under certain conditions. The degree of the equivalent model is intricately linked to both the "data spread" and the learning rate employed during one gradient step. Through extensive simulations, we demonstrate the equivalence between the original model and its polynomial counterpart across various regression and classification tasks. Additionally, we explore how different properties of Gaussian mixtures affect learning outcomes. Finally, we illustrate experimental results on Fashion-MNIST classification, indicating that our findings can translate to realistic data., Comment: ICLR 2025, 27 pages, 9 figures

Hybrid paper interfaces leverage augmented reality to combine the desired tangibility of paper documents with the affordances of interactive digital media. Typically, virtual content can be embedded through direct links (e.g., QR codes); however, this impacts the aesthetics of the paper print and limits the available visual content space. To address this problem, we present Imprinto, an infrared inkjet watermarking technique that allows for invisible content embeddings only by using off-the-shelf IR inks and a camera. Imprinto was established through a psychophysical experiment, studying how much IR ink can be used while remaining invisible to users regardless of background color. We demonstrate that we can detect invisible IR content through our machine learning pipeline, and we developed an authoring tool that optimizes the amount of IR ink on the color regions of an input document for machine and human detectability. Finally, we demonstrate several applications, including augmenting paper documents and objects., Comment: 18 pages, 13 figures. To appear in the Proceedings of the 2025 ACM CHI Conference on Human Factors in Computing Systems. https://imprinto.github.io

X-ray fluorescence computed tomography (XFCT), a form of X-ray molecular imaging, offers detailed quantitative imaging capabilities for high-Z metal nanoparticles (MNPs), which are widely studied for their applications in multifunctional theranostics. Due to its affordability and accessibility, the benchtop XFCT prototype typically employs a single-pixel detector (SPD) with single-pencil-beam (SPB) X-ray excitation. While this design (resembling the first-generation CT geometry) achieves reliable detection sensitivity, it is hindered by long imaging times. The use of simultaneous multiple-pencil-beam (MPB) excitation presents a promising solution to significantly reduce imaging times. In this study, we developed a repeatable workflow that combines Monte Carlo (MC) simulations and 3D printing to design Nbeam-MPB collimator, where Nbeam is the number of beams generated by the collimator. As an initial test, we fabricated a 2-MPB collimator and evaluated the performance of 2-MPB-based XFCT imaging on a physical phantom and small animals surgically implanted with agarose pellets containing gold chloride (H[AuCl4]). The results demonstrated a 2x acceleration in image acquisition without compromising the contrast-to-noise ratio (CNR). We further investigated the concept of Nbeam-MPB acceleration on the MC computational XFCT system, which confirmed the feasibility of achieving at least 4x acceleration with 4-MPB excitation. Combined with additional system optimization, such as X-ray beam flux optimization, XFCT imaging could be further accelerated, reducing acquisition time from hours to minutes and meeting the requirements for routine MNP imaging.

This article investigates crime patterns across European countries in 2022 using Compositional Data Analysis (CoDA) to address limitations of traditional statistical approaches in dealing with the relative nature of crime data. Recognizing crime types as components of a whole, we employ CoDA to explore relationships between different crime categories while respecting their inherent interdependencies. The study utilizes k-means clustering to group countries based on their crime profiles, identifying three distinct clusters largely aligning with geographical locations. This clustering is visualized through t-SNE and geographic mapping, revealing regional similarities. Further analysis using Robust Principal Component Analysis on identified crime clusters reveals insightful relationships between specific crime types, such as homicide, smuggling, and financial crimes, and how their prevalence varies across countries. The findings reveals distinct crime patterns across Europe, highlighting regional commonalities while also highlighting divergences like Norway and Latvia that deviate from their expected geographical classifications. Moreover, the study identifies specific crime groups; for example, it pairs countries high in corruption and smuggling, such as Austria, with those countries that exhibit a higher relevance to homicide and smuggling, such as Luxembourg. It also points to the presence of financial crimes like fraud in countries such as Romania and Estonia.

Creating new fonts requires a lot of human effort and professional typographic knowledge. Despite the rapid advancements of automatic font generation models, existing methods require users to prepare pre-designed characters with target styles using font-editing software, which poses a problem for non-expert users. To address this limitation, we propose FontCraft, a system that enables font generation without relying on pre-designed characters. Our approach integrates the exploration of a font-style latent space with human-in-the-loop preferential Bayesian optimization and multimodal references, facilitating efficient exploration and enhancing user control. Moreover, FontCraft allows users to revisit previous designs, retracting their earlier choices in the preferential Bayesian optimization process. Once users finish editing the style of a selected character, they can propagate it to the remaining characters and further refine them as needed. The system then generates a complete outline font in OpenType format. We evaluated the effectiveness of FontCraft through a user study comparing it to a baseline interface. Results from both quantitative and qualitative evaluations demonstrate that FontCraft enables non-expert users to design fonts efficiently., Comment: 14 pages

We analyze a non-supersymmetric, non-renormalizable grand unified theory whose particle content is that of the Georgi-Glashow model augmented only by scalars from the \textbf{10} and \textbf{35} representations. A prediction of our model is a color sextet, weak isodoublet whose mass lies at 1 TeV (10 TeV) and that does not couple to Standard Model fermions at tree level. The leptoquark mechanism through which Majorana neutrinos radiatively acquire their masses relates the neutrino mass matrix to that of the down-type quarks. A consequence of this relation and perturbativity of coupling constants is the upper bound of $2.5 \times 10^{15}$ GeV on the masses of the scalar leptoquarks $S_1^*$ and ${\widetilde R}_{2}$. Electroweak mixing of these leptoquarks induces a B-L violating decay of the proton which indirectly contrains the mass of ${\widetilde R}_{2}$ to be greater than $2.1 \times 10^9$ GeV. We find the grand unification scale to exceed $1.4 \times 10^{16}$ GeV in all scenarios considered.

We compute the conservative and radiation-reaction contributions to classical observables in the gravitational scattering between a spinning and a spinless black hole to the fourth order in spin and third order in the gravitational constant. The conservative results are obtained from two-loop amplitudes for the scattering process of a massive scalar with a massive spin-$s$ field $(s=0, 1, 2)$ minimally coupled to gravity, employing the recently introduced spin interpolation method to resolve all spin-Casimir terms. The two-loop amplitude exhibits a spin-shift symmetry in both probe limits, which we conjecture to be a sign of yet unknown integrability of Kerr orbits through the quartic order in spin and to all orders in the gravitational constant. We obtain the radial action from the finite part of the amplitude and use it to compute classical observables, including the impulse and spin kick. This is done using the recently introduced covariant Dirac brackets, which allow for the computation of classical scattering observables for general (non-aligned) spin configurations. Finally, employing the radiation-reaction amplitude proposed by Alessio and Di Vecchia, together with the Dirac brackets, we obtain radiation-reaction contributions to observables at all orders in spin and beyond the aligned-spin limit. We find agreement with known results up to the quadratic order in spin for both conservative and radiation-reaction contributions. Our results advance the state of the art in the understanding of spinning binary dynamics in general relativity and demonstrate the power and simplicity of the Dirac bracket formalism for relating scattering amplitudes to classical observables., Comment: 13 pages; v2: adds references, corrects observables in ancillary files

Research has demonstrated that zeolite nucleation and growth can be controlled by fine-tuning chemical composition, temperature, and pressure, resulting in structures with diverse porosities and functionalities. Nevertheless, current energy delivery methods lack the finesse required to operate on the femto- and picosecond timescales of silica polymerisation and depolymerisation, limiting their ability to direct synthesis with high precision. To overcome this limitation, we introduce an ultrafast laser synthesis technique capable of delivering energy at these timescales with unprecedented spatiotemporal precision. Unlike conventional or emerging approaches, this method bypasses the need for specific temperature and pressure settings, as nucleation and growth are governed by dynamic phenomena arising from nonlinear light-matter interactions, such as convective flows, cavitation bubbles, plasma formation, and shock waves. These processes can be initiated, paused, and resumed within fractions of a second, effectively freezing structures at any stage of self-assembly. Using this approach, we traced the entire nucleation and growth pathway of laser-synthesized TPA-silicate-1 zeolites, from early oligomer formation to fully developed crystals. The unprecedented spatiotemporal control of this technique unlocks new avenues for manipulating reaction pathways and exploring the vast configurational space of zeolites., Comment: main text, 6 figures, 4 video files, supplementary information

Creating custom artifacts with computer numerical control (CNC) milling machines typically requires mastery of complex computer-aided design (CAD) software. To eliminate this user barrier, we introduced Draw2Cut, a novel system that allows users to design and fabricate artifacts by sketching directly on physical materials. Draw2Cut employs a custom-drawing language to convert user-drawn lines, symbols, and colors into toolpaths, thereby enabling users to express their creative intent intuitively. The key features include real-time alignment between material and virtual toolpaths, a preview interface for validation, and an open-source platform for customization. Through technical evaluations and user studies, we demonstrate that Draw2Cut lowers the entry barrier for personal fabrication, enabling novices to create customized artifacts with precision and ease. Our findings highlight the potential of the system to enhance creativity, engagement, and accessibility in CNC-based woodworking., Comment: Accepted by CHI 2025

For many shape analysis problems in computer vision and scientific imaging (e.g., computational anatomy, morphological cytometry), the ability to align two closed curves in the plane is crucial. In this paper, we concentrate on rigidly aligning pairs of closed curves in the plane. If the curves have the same length and are centered at the origin, the critical steps to an optimal rigid alignment are finding the best rotation for one curve to match the other and redefining the starting point of the rotated curve so that the starting points of the two curves match. Unlike open curves, closed curves do not have fixed starting points, and this introduces an additional degree of freedom in the alignment. Hence the common naive method to find the best rotation and starting point for optimal rigid alignment has O(N^2) time complexity, N the number of nodes per curve. This can be slow for curves with large numbers of nodes. In this paper, we propose a new O(N log N) algorithm for this problem based on the Fast Fourier Transform. Together with uniform resampling of the curves with respect to arc length, the new algorithm results in an order of magnitude speed-up in our experiments. Additionally, we describe how we can use our new algorithm as part of elastic shape distance computations between closed curves to obtain accurate shape distance values at a fraction of the cost of previous approaches.

Woodworkers have to navigate multiple considerations when planning a project, including available resources, skill-level, and intended effort. Do it yourself (DIY) woodworkers face these challenges most acutely because of tight material constraints and a desire for custom designs tailored to specific spaces. To address these needs, we present XR-penter, an extended reality (XR) application that supports in situ, material-aware woodworking for casual makers. Our system enables users to design virtual scrap wood assemblies directly in their workspace, encouraging sustainable practices through the use of discarded materials. Users register physical material as virtual twins, manipulate these twins into an assembly in XR, and preview cuts needed for fabrication. We conducted a case study and feedback sessions to demonstrate how XR-penter supports improvisational workflows in practice, the type of woodworker who would benefit most from our system, and insights on integrating similar spatial and material considerations into future work.

This paper introduces foundational resources and models for natural language processing (NLP) of historical Turkish, a domain that has remained underexplored in computational linguistics. We present the first named entity recognition (NER) dataset, HisTR and the first Universal Dependencies treebank, OTA-BOUN for a historical form of the Turkish language along with transformer-based models trained using these datasets for named entity recognition, dependency parsing, and part-of-speech tagging tasks. Additionally, we introduce Ottoman Text Corpus (OTC), a clean corpus of transliterated historical Turkish texts that spans a wide range of historical periods. Our experimental results show significant improvements in the computational analysis of historical Turkish, achieving promising results in tasks that require understanding of historical linguistic structures. They also highlight existing challenges, such as domain adaptation and language variations across time periods. All of the presented resources and models are made available at https://huggingface.co/bucolin to serve as a benchmark for future progress in historical Turkish NLP.

We propose a novel method for the efficient and accurate iterative solution of frequency domain integral equations (IEs) that are used for large/multi-scale electromagnetic scattering problems. The proposed method uses a novel group-by-group interaction strategy to accurately evaluate far-zone interactions within the framework of the one-box-buffer scheme during the matrix-vector multiplication at each iteration. Briefly, subdomain basis functions that are used to model the scatterer at each box are represented by a fixed number of uniformly distributed and arbitrarily oriented Hertzian dipoles (referred to as uniform basis functions), and then the dipole-to-dipole interactions are predicted in a group-wise manner by employing machine learning algorithms, thereby showcasing efficiency, strong scalability for parallelization and accuracy without the low-frequency breakdown (LFB) problem. Since the dipole representation is independent of the underlying material properties of the scatterer, the proposed method is valid for all types of IEs (surface or volume). Moreover, because the training is performed offline, the resulting networks can be used for any scatterer under any IE, without extra training, as long as the size of, and the distances among the boxes are preserved. The efficiency and accuracy of the proposed method are assessed by comparing our results with those obtained from the conventional multilevel fast multipole algorithm for various scattering problems. The proposed method's parallelization performance is showcased through scalability tests, and its resilience to LFB is demonstrated., Comment: 18 pages, 16 figures, 1 tables. Submitted for publication and is under review in IEEE Transactions on Antennas and Propagation, 2024

This study investigates the behavior of photonic modes in twisted graphene nanoribbons (TGNRs) using an analytical approach based on solving the fully covariant vector boson equation. We present a model that demonstrates how helical twisting in TGNRs significantly affects the evolution of photonic modes. Our analytical solutions yield detailed expressions for mode profiles, energy spectra, and decay characteristics. We find that increasing the twist parameter shortens the decay times (\(\tau_{ns}\)) for damped modes, indicating enhanced photonic coupling due to the twisted geometry. Conversely, longer nanoribbons (NRs) exhibit increased decay times, showing a length (\(L\))-dependent effect, where \(\tau_{ns} \propto L / c\), with \(c\) representing the speed of light. These findings may enhance the understanding of light control in nanostructures and suggest potential applications in tunable photonic devices, topological photonics, and quantum optical systems., Comment: 5 pages, 1 figure

We study compactness of Hankel and Toeplitz operators on Bergman spaces of convex Reinhardt domains in $\mathbb{C}^2$ and we restrict the symbols to the class of functions that are continuous on the closure of the domain. We prove that Toeplitz operators as well as the Hermitian squares of Hankel operators are compact if and only if the Berezin transforms of the operators vanish on the boundary of the domain.

The design and performance analysis of relay lenses that provide high-performance image transmission for target acquisition and tracking in military optical systems. Relay lenses are critical components for clear and lossless image transmission over long distances. In this study, the optical performance of a relay lens system designed and optimized using ZEMAX software is investigated in detail. The analysis focuses on important optical properties such as modulation transfer function (MTF), spot diagrams, Seidel diagram, field curvature and distortion. The results show that the lens has significant potential in military applications for target detection and tracking with high resolution and low aberration.

There is growing concern about our vulnerability to space weather hazards and the disruption critical infrastructure failures could cause to society and the economy. However, the socio-economic impacts of space weather hazards, such as from geomagnetic storms, remain under-researched. This study introduces a novel framework to estimate the economic impacts of electricity transmission infrastructure failure due to space weather. By integrating existing geophysical and geomagnetically induced current (GIC) estimation models with a newly developed geospatial model of the Continental United States power grid, GIC vulnerabilities are assessed for a range of space weather scenarios. The approach evaluates multiple power network architectures, incorporating input-output economic modeling to translate business and population disruptions into macroeconomic impacts from GIC-related thermal heating failures. The results indicate a daily GDP loss from 6 billion USD to over 10 billion USD. Even under conservative GIC thresholds (75 A/ph) aligned with thermal withstand limits from the North American Electric Reliability Corporation (NERC), significant economic disruptions are evident. This study is limited by its restriction to thermal heating analysis, though GICs can also affect the grid through other pathways, such as voltage instability and harmonic distortions. Addressing these other failure mechanisms need to be the focus of future research.

Creating high-quality 3D avatars using 3D Gaussian Splatting (3DGS) from a monocular video benefits virtual reality and telecommunication applications. However, existing automatic methods exhibit artifacts under novel poses due to limited information in the input video. We propose AvatarPerfect, a novel system that allows users to iteratively refine 3DGS avatars by manually editing the rendered avatar images. In each iteration, our system suggests a new body and camera pose to help users identify and correct artifacts. The edited images are then used to update the current avatar, and our system suggests the next body and camera pose for further refinement. To investigate the effectiveness of AvatarPerfect, we conducted a user study comparing our method to an existing 3DGS editor SuperSplat, which allows direct manipulation of Gaussians without automatic pose suggestions. The results indicate that our system enables users to obtain higher quality refined 3DGS avatars than the existing 3DGS editor., Comment: 13 pages

In this study, we develop and assess new corpus selection and training methodologies to improve the effectiveness of Turkish language models. Specifically, we adapted Large Language Model generated datasets and translated English datasets into Turkish, integrating these resources into the training process. This approach led to substantial enhancements in model accuracy for both few-shot and zero-shot learning scenarios. Furthermore, the merging of these adapted models was found to markedly improve their performance. Human evaluative metrics, including task-specific performance assessments, further demonstrated that these adapted models possess a greater aptitude for comprehending the Turkish language and addressing logic-based queries. This research underscores the importance of refining corpus selection strategies to optimize the performance of multilingual models, particularly for under-resourced languages like Turkish., Comment: 2024 Innovations in Intelligent Systems and Applications Conference (ASYU)

In this study, a Turkish visual instruction model was developed and various model architectures and dataset combinations were analysed to improve the performance of this model. The Cosmos-LLaVA model, which is built by combining different large language models and image coders, is designed to overcome the deficiencies in the Turkish language. In the experiments, the effects of fine-tuning with various datasets on the model performance are analysed in detail. The results show that model architecture and dataset selection have a significant impact on performance. Bu \c{c}al{\i}\c{s}mada bir T\"urk\c{c}e g\"orsel talimat modeli geli\c{s}tirilerek bu modelin performans{\i}n{\i} art{\i}rmaya y\"onelik \c{c}e\c{s}itli model mimarileri ve veri k\"umesi kombinasyonlar{\i} derinlemesine incelenmi\c{s}tir. Farkl{\i} b\"uy\"uk dil modelleri ve g\"or\"unt\"u kodlay{\i}c{\i}lar{\i}n{\i}n bir araya getirilmesiyle olu\c{s}turulan Cosmos-LLaVA modeli, T\"urk\c{c}e dilindeki eksiklikleri gidermeye y\"onelik olarak tasarlanm{\i}\c{s}t{\i}r. Yap{\i}lan deneylerde, \c{c}e\c{s}itli veri k\"umeleri ile yap{\i}lan ince ayarlar{\i}n model performans{\i}n{\i} nas{\i}l etkiledi\u{g}i detayl{\i} olarak ele al{\i}nm{\i}\c{s}t{\i}r. Sonu\c{c}lar, model mimarisi ve veri k\"umesi se\c{c}iminin performans \"uzerinde \"onemli bir etkiye sahip oldu\u{g}unu g\"ostermektedir., Comment: in Turkish language, 2024 8th International Artificial Intelligence and Data Processing Symposium (IDAP)

Surgical treatment for prostate cancer often involves organ removal, i.e., prostatectomy. Pathology reports on these specimens convey treatment-relevant information. Beyond these reports, the diagnostic process generates extensive and complex information that is difficult to represent in reports, although it is of significant interest to the other medical specialties involved. 3D tissue reconstruction would allow for better spatial visualization, as well as combinations with other imaging modalities. Existing approaches in this area have proven labor-intensive and challenging to integrate into clinical workflows. 3D-SLIVER provides a simplified solution, implemented as an open-source 3DSlicer extension. We outline three specific real-world scenarios to illustrate its potential to improve transparency in diagnostic workflows and contribute to multi-modal research endeavors. Implementing the 3D reconstruction process involved four sub-modules of 3D-SLIVER: digitization of slicing protocol, virtual slicing of arbitrary 3D models based on that protocol, registration of slides with virtual slices using the Coherent Point Drift algorithm, and 3D reconstruction of registered information using convex hulls, Gaussian splatter and linear extrusion. Three use cases to employ 3D-SLIVER are presented: a low-effort approach to pathology workflow integration and two research-related use cases illustrating how to perform retrospective evaluations of PI-RADS predictions and statistically model 3D distributions of morphological patterns. 3D-SLIVER allows for improved interdisciplinary communication among specialties. It is designed for simplicity in application, allowing for flexible integration into various workflows and use cases. Here we focused on the clinical care of prostate cancer patients, but future possibilities are extensive with other neoplasms and in education and research.

Wikidata has a large ontology with classes at several orders. The Wikidata ontology has long been known to have violations of class order and information related to class order that appears suspect. SPARQL queries were evaluated against Wikidata to determine the prevalence of several kinds of violations and suspect information and the results analyzed. Some changes were manually made to Wikidata to remove some of these results and the queries rerun, showing the effect of the changes. Suggestions are provided on how the problems uncovered might be addressed, either though better tooling or involvement of the Wikidata community.

In this study, we propose HOPER (HOlistic ProtEin Representation), a novel multimodal learning framework designed to enhance protein function prediction (PFP) in low-data settings. The challenge of predicting protein functions is compounded by the limited availability of labeled data. Traditional machine learning models already struggle in such cases, and while deep learning models excel with abundant data, they also face difficulties when data is scarce. HOPER addresses this issue by integrating three distinct modalities - protein sequences, biomedical text, and protein-protein interaction (PPI) networks - to create a comprehensive protein representation. The model utilizes autoencoders to generate holistic embeddings, which are then employed for PFP tasks using transfer learning. HOPER outperforms existing methods on a benchmark dataset across all Gene Ontology categories, i.e., molecular function, biological process, and cellular component. Additionally, we demonstrate its practical utility by identifying new immune-escape proteins in lung adenocarcinoma, offering insights into potential therapeutic targets. Our results highlight the effectiveness of multimodal representation learning for overcoming data limitations in biological research, potentially enabling more accurate and scalable protein function prediction. HOPER source code and datasets are available at https://github.com/kansil/HOPER

In most applications, robots need to adapt to new environments and be multi-functional without forgetting previous information. This requirement gains further importance in real-world scenarios where robots operate in coexistence with humans. In these complex environments, human actions inevitably lead to changes, requiring robots to adapt accordingly. To effectively address these dynamics, the concept of continual learning proves essential. It not only enables learning models to integrate new knowledge while preserving existing information but also facilitates the acquisition of insights from diverse contexts. This aspect is particularly relevant to the issue of context-switching, where robots must navigate and adapt to changing situational dynamics. Our approach introduces a novel approach to effectively tackle the problem of context drifts by designing a Streaming Graph Neural Network that incorporates both regularization and rehearsal techniques. Our Continual\_GTM model enables us to retain previous knowledge from different contexts, and it is more effective than traditional fine-tuning approaches. We evaluated the efficacy of Continual\_GTM in predicting human routines within household environments, leveraging spatio-temporal object dynamics across diverse scenarios.

Neural contextual biasing allows speech recognition models to leverage contextually relevant information, leading to improved transcription accuracy. However, the biasing mechanism is typically based on a cross-attention module between the audio and a catalogue of biasing entries, which means computational complexity can pose severe practical limitations on the size of the biasing catalogue and consequently on accuracy improvements. This work proposes an approximation to cross-attention scoring based on vector quantization and enables compute- and memory-efficient use of large biasing catalogues. We propose to use this technique jointly with a retrieval based contextual biasing approach. First, we use an efficient quantized retrieval module to shortlist biasing entries by grounding them on audio. Then we use retrieved entries for biasing. Since the proposed approach is agnostic to the biasing method, we investigate using full cross-attention, LLM prompting, and a combination of the two. We show that retrieval based shortlisting allows the system to efficiently leverage biasing catalogues of several thousands of entries, resulting in up to 71% relative error rate reduction in personal entity recognition. At the same time, the proposed approximation algorithm reduces compute time by 20% and memory usage by 85-95%, for lists of up to one million entries, when compared to standard dot-product cross-attention., Comment: 13 pages, 7 figures, submitted to IEEE/ACM Transactions on Audio, Speech, and Language Processing

Subspace learning is a critical endeavor in contemporary machine learning, particularly given the vast dimensions of modern datasets. In this study, we delve into the training dynamics of a single-layer GAN model from the perspective of subspace learning, framing these GANs as a novel approach to this fundamental task. Through a rigorous scaling limit analysis, we offer insights into the behavior of this model. Extending beyond prior research that primarily focused on sequential feature learning, we investigate the non-sequential scenario, emphasizing the pivotal role of inter-feature interactions in expediting training and enhancing performance, particularly with an uninformed initialization strategy. Our investigation encompasses both synthetic and real-world datasets, such as MNIST and Olivetti Faces, demonstrating the robustness and applicability of our findings to practical scenarios. By bridging our analysis to the realm of subspace learning, we systematically compare the efficacy of GAN-based methods against conventional approaches, both theoretically and empirically. Notably, our results unveil that while all methodologies successfully capture the underlying subspace, GANs exhibit a remarkable capability to acquire a more informative basis, owing to their intrinsic ability to generate new data samples. This elucidates the unique advantage of GAN-based approaches in subspace learning tasks., Comment: Accepted for NeurIPS 2024, 16 pages, 7 figures

We introduce a dynamic spatiotemporal volatility model that extends traditional approaches by incorporating spatial, temporal, and spatiotemporal spillover effects, along with volatility-specific observed and latent factors. The model offers a more general network interpretation, making it applicable for studying various types of network spillovers. The primary innovation lies in incorporating volatility-specific latent factors into the dynamic spatiotemporal volatility model. Using Bayesian estimation via the Markov Chain Monte Carlo (MCMC) method, the model offers a robust framework for analyzing the spatial, temporal, and spatiotemporal effects of a log-squared outcome variable on its volatility. We recommend using the deviance information criterion (DIC) and a regularized Bayesian MCMC method to select the number of relevant factors in the model. The model's flexibility is demonstrated through two applications: a spatiotemporal model applied to the U.S. housing market and another applied to financial stock market networks, both highlighting the model's ability to capture varying degrees of interconnectedness. In both applications, we find strong spatial/network interactions with relatively stronger spillover effects in the stock market.

One of the fundamental problems in computer vision is image segmentation, the task of detecting distinct regions or objects in given images. Deep Neural Networks (DNN) have been shown to be very effective in segmenting challenging images, producing convincing segmentations. There is further need for probabilistic DNNs that can reflect the uncertainties from the input images and the models into the computed segmentations, in other words, new DNNs that can generate multiple plausible segmentations and their distributions depending on the input or the model uncertainties. While there are existing probabilistic segmentation models, many of them do not take into account the geometry or shape underlying the segmented regions. In this paper, we propose a probabilistic image segmentation model that can incorporate the geometry of a segmentation. Our proposed model builds on the Probabilistic U-Net of \cite{kohl2018probabilistic} to generate probabilistic segmentations, i.e.\! multiple likely segmentations for an input image. Our model also adopts the Kendall Shape Variational Auto-Encoder of \cite{vadgama2023kendall} to encode a Kendall shape space in the latent variable layers of the prior and posterior networks of the Probabilistic U-Net. Incorporating the shape space in this manner leads to a more robust segmentation with spatially coherent regions, respecting the underlying geometry in the input images., Comment: 22 pages, 13 figures

Disjointness checks are among the most important constraint checks in a knowledge base and can be used to help detect and correct incorrect statements and internal contradictions. Wikidata is a very large, community-managed knowledge base. Because of both its size and construction, Wikidata contains many incorrect statements and internal contradictions. We analyze the current modeling of disjointness on Wikidata, identify patterns that cause these disjointness violations and categorize them. We use SPARQL queries to identify each ``culprit'' causing a disjointness violation and lay out formulas to identify and fix conflicting information. We finally discuss how disjointness information could be better modeled and expanded in Wikidata in the future., Comment: Sixth International Knowledge Graph and Semantic Web Conference

This paper introduces Robotic Augmented Reality for Machine Programming by Demonstration (RAMPA), the first ML-integrated, XR-driven end-to-end robotic system, allowing training and deployment of ML models such as ProMPs on the fly, and utilizing the capabilities of state-of-the-art and commercially available AR headsets, e.g., Meta Quest 3, to facilitate the application of Programming by Demonstration (PbD) approaches on industrial robotic arms, e.g., Universal Robots UR10. Our approach enables in-situ data recording, visualization, and fine-tuning of skill demonstrations directly within the user's physical environment. RAMPA addresses critical challenges of PbD, such as safety concerns, programming barriers, and the inefficiency of collecting demonstrations on the actual hardware. The performance of our system is evaluated against the traditional method of kinesthetic control in teaching three different robotic manipulation tasks and analyzed with quantitative metrics, measuring task performance and completion time, trajectory smoothness, system usability, user experience, and task load using standardized surveys. Our findings indicate a substantial advancement in how robotic tasks are taught and refined, promising improvements in operational safety, efficiency, and user engagement in robotic programming., Comment: This work is the final version submitted to the IEEE RA-L

Robust classification of the operational environment of wireless devices is becoming increasingly important for wireless network optimization, particularly in a shared spectrum environment. Distinguishing between indoor and outdoor devices can enhance reliability and improve coexistence with existing, outdoor, incumbents. For instance, the unlicensed but shared 6 GHz band (5.925 - 7.125 GHz) enables sharing by imposing lower transmit power for indoor unlicensed devices and a spectrum coordination requirement for outdoor devices. Further, indoor devices are prohibited from using battery power, external antennas, and weatherization to prevent outdoor operations. As these rules may be circumvented, we propose a robust indoor/outdoor classification method by leveraging the fact that the radio-frequency environment faced by a device are quite different indoors and outdoors. We first collect signal strength data from all cellular and Wi-Fi bands that can be received by a smartphone in various environments (indoor interior, indoor near windows, and outdoors), along with GPS accuracy, and then evaluate three machine learning (ML) methods: deep neural network (DNN), decision tree, and random forest to perform classification into these three categories. Our results indicate that the DNN model performs the best, particularly in minimizing the most important classification error, that of classifying outdoor devices as indoor interior devices.

Materials scientists utilize image segmentation of micrographs to create large curve ensembles representing grain boundaries of material microstructures. Observations of these collections of shapes can facilitate inferences about material properties and manufacturing processes. We seek to bolster this application, and related engineering/scientific tasks, using novel pattern recognition formalisms and inference over large ensembles of segmented curves -- i.e., facilitate principled assessments for quantifying differences in distributions of shapes. To this end, we apply a composite integral operator to motivate accurate and efficient numerical representations of discrete planar curves over matrix manifolds. The main result is a rigid-invariant orthonormal decomposition of curve component functions into separable forms of scale variations and complementary features of undulation. We demonstrate how these separable shape tensors -- given thousands of curves in an ensemble -- can inform explainable binary classification of segmented images by utilizing a product maximum mean discrepancy to distinguish the shape distributions; absent labelled data, building interpretable feature spaces in seconds without high performance computation, and detecting discrepancies below cursory visual inspections., Comment: 20 pages, 7 figures

The objective of this work is to develop an AI foundation model for physical signals that can generalize across diverse phenomena, domains, applications, and sensing apparatuses. We propose a phenomenological approach and framework for creating and validating such AI foundation models. Based on this framework, we developed and trained a model on 0.59 billion samples of cross-modal sensor measurements, ranging from electrical current to fluid flow to optical sensors. Notably, no prior knowledge of physical laws or inductive biases were introduced into the model. Through several real-world experiments, we demonstrate that a single foundation model could effectively encode and predict physical behaviors, such as mechanical motion and thermodynamics, including phenomena not seen in training. The model also scales across physical processes of varying complexity, from tracking the trajectory of a simple spring-mass system to forecasting large electrical grid dynamics. This work highlights the potential of building a unified AI foundation model for diverse physical world processes.

Despite the growing accessibility of augmented reality (AR) for visualization, existing computer-aided design (CAD) systems remain confined to traditional screens or require complex setups or predefined parameters, limiting immersion and accessibility for novices. We present TinkerXR, an open-sourced interface enabling in-situ design and fabrication through Constructive Solid Geometry (CSG) modeling. TinkerXR operates solely with a headset and 3D printer, allowing users to design directly in and for their physical environments. By leveraging spatial awareness, depth occlusion, recognition of physical constraints, reference objects, and intuitive hand movement controls, TinkerXR enhances realism, precision, and ease of use. Its AR-based workflow integrates design and 3D printing with a drag-and-drop interface for a 3D printer's virtual twin. A user study comparing TinkerXR with Tinkercad demonstrates higher accessibility, engagement, and ease of use for novices. By bridging the gap between digital creation and physical output, TinkerXR transforms everyday spaces into accessible and expressive creative studios., Comment: 28 pages, 18 figures, 1 table

Precision pose detection is increasingly demanded in fields such as personal fabrication, Virtual Reality (VR), and robotics due to its critical role in ensuring accurate positioning information. However, conventional vision-based systems used in these systems often struggle with achieving high precision and accuracy, particularly when dealing with complex environments or fast-moving objects. To address these limitations, we investigate Laser Speckle Imaging (LSI), an emerging optical tracking method that offers promising potential for improving pose estimation accuracy. Specifically, our proposed LSI-Based Tracking (SpecTrack) leverages the captures from a lensless camera and a retro-reflector marker with a coded aperture to achieve multi-axis rotational pose estimation with high precision. Our extensive trials using our in-house built testbed have shown that SpecTrack achieves an accuracy of 0.31{\deg} (std=0.43{\deg}), significantly outperforming state-of-the-art approaches and improving accuracy up to 200%.