Search

Self-supervised learning has recently emerged as the preeminent pretraining paradigm across and between modalities, with remarkable results. In the image domain specifically, group (or cluster) discrimination has been one of the most successful methods. However, such frameworks need to guard against heavily imbalanced cluster assignments to prevent collapse to trivial solutions. Existing works typically solve this by reweighing cluster assignments to promote balance, or with offline operations (e.g. regular re-clustering) that prevent collapse. However, the former typically requires large batch sizes, which leads to increased resource requirements, and the latter introduces scalability issues with regard to large datasets. In this work, we propose ExCB, a framework that tackles this problem with a novel cluster balancing method. ExCB estimates the relative size of the clusters across batches and balances them by adjusting cluster assignments, proportionately to their relative size and in an online manner. Thereby, it overcomes previous methods' dependence on large batch sizes and is fully online, and therefore scalable to any dataset. We conduct extensive experiments to evaluate our approach and demonstrate that ExCB: a) achieves state-of-the-art results with significantly reduced resource requirements compared to previous works, b) is fully online, and therefore scalable to large datasets, and c) is stable and effective even with very small batch sizes., Comment: Accepted at ECCV 2024

We present multiwavelength polarization measurements of the luminous blazar Mrk~501 over a 14-month period. The 2--8 keV X-ray polarization was measured with the Imaging X-ray Polarimetry Explorer (IXPE) with six 100-ks observations spanning from 2022 March to 2023 April. Each IXPE observation was accompanied by simultaneous X-ray data from NuSTAR, Swift/XRT, and/or XMM-Newton. Complementary optical-infrared polarization measurements were also available in the B, V, R, I, and J bands, as were radio polarization measurements from 4.85 GHz to 225.5 GHz. Among the first five IXPE observations, we did not find significant variability in the X-ray polarization degree and angle with IXPE. However, the most recent sixth observation found an elevated polarization degree at $>3\sigma$ above the average of the other five observations. The optical and radio measurements show no apparent correlations with the X-ray polarization properties. Throughout the six IXPE observations, the X-ray polarization degree remained higher than, or similar to, the R-band optical polarization degree, which remained higher than the radio value. This is consistent with the energy-stratified shock scenario proposed to explain the first two IXPE observations, in which the polarized X-ray, optical, and radio emission arises from different regions., Comment: Accepted for publication in ApJ

In this paper, we extend the theory of planar pseudo knots to the theories of annular and toroidal pseudo knots. Pseudo knots are defined as equivalence classes under Reidemeister-like moves of knot diagrams characterized by crossings with undefined over/under information. In the theories of annular and toroidal pseudo knots we introduce their respective lifts to the solid and the thickened torus. Then, we interlink these theories by representing annular and toroidal pseudo knots as planar ${\rm O}$-mixed and ${\rm H}$-mixed pseudo links. We also explore the inclusion relations between planar, annular and toroidal pseudo knots, as well as of ${\rm O}$-mixed and ${\rm H}$-mixed pseudo links. Finally, we extend the planar weighted resolution set to annular and toroidal pseudo knots, defining new invariants for classifying pseudo knots and links in the solid and in the thickened torus., Comment: 22 pages, 20 figures

This paper studies planar drawings of graphs in which each vertex is represented as a point along a sequence of horizontal lines, called levels, and each edge is either a horizontal segment or a strictly $y$-monotone curve. A graph is $s$-span weakly leveled planar if it admits such a drawing where the edges have span at most $s$; the span of an edge is the number of levels it touches minus one. We investigate the problem of computing $s$-span weakly leveled planar drawings from both the computational and the combinatorial perspectives. We prove the problem to be para-NP-hard with respect to its natural parameter $s$ and investigate its complexity with respect to widely used structural parameters. We show the existence of a polynomial-size kernel with respect to vertex cover number and prove that the problem is FPT when parameterized by treedepth. We also present upper and lower bounds on the span for various graph classes. Notably, we show that cycle trees, a family of $2$-outerplanar graphs generalizing Halin graphs, are $\Theta(\log n)$-span weakly leveled planar and $4$-span weakly leveled planar when $3$-connected. As a byproduct of these combinatorial results, we obtain improved bounds on the edge-length ratio of the graph families under consideration., Comment: Appears in the Proceedings of the 32nd International Symposium on Graph Drawing and Network Visualization (GD 2024)

In visual decision making, high-level features, such as object categories, have a strong influence on choice. However, the impact of low-level features on behavior is less understood partly due to the high correlation between high- and low-level features in the stimuli presented (e.g., objects of the same category are more likely to share low-level features). To disentangle these effects, we propose a method that de-correlates low- and high-level visual properties in a novel set of stimuli. Our method uses two Convolutional Neural Networks (CNNs) as candidate models of the ventral visual stream: the CORnet-S that has high neural predictivity in high-level, IT-like responses and the VGG-16 that has high neural predictivity in low-level responses. Triplets (root, image1, image2) of stimuli are parametrized by the level of low- and high-level similarity of images extracted from the different layers. These stimuli are then used in a decision-making task where participants are tasked to choose the most similar-to-the-root image. We found that different networks show differing abilities to predict the effects of low-versus-high-level similarity: while CORnet-S outperforms VGG-16 in explaining human choices based on high-level similarity, VGG-16 outperforms CORnet-S in explaining human choices based on low-level similarity. Using Brain-Score, we observed that the behavioral prediction abilities of different layers of these networks qualitatively corresponded to their ability to explain neural activity at different levels of the visual hierarchy. In summary, our algorithm for stimulus set generation enables the study of how different representations in the visual stream affect high-level cognitive behaviors., Comment: Accepted at Workshop on Human-inspired Computer Vision @ ECCV2024

The relentless expansion of deep learning applications in recent years has prompted a pivotal shift toward on-device execution, driven by the urgent need for real-time processing, heightened privacy concerns, and reduced latency across diverse domains. This article addresses the challenges inherent in optimising the execution of deep neural networks (DNNs) on mobile devices, with a focus on device heterogeneity, multi-DNN execution, and dynamic runtime adaptation. We introduce CARIn, a novel framework designed for the optimised deployment of both single- and multi-DNN applications under user-defined service-level objectives. Leveraging an expressive multi-objective optimisation framework and a runtime-aware sorting and search algorithm (RASS) as the MOO solver, CARIn facilitates efficient adaptation to dynamic conditions while addressing resource contention issues associated with multi-DNN execution. Notably, RASS generates a set of configurations, anticipating subsequent runtime adaptation, ensuring rapid, low-overhead adjustments in response to environmental fluctuations. Extensive evaluation across diverse tasks, including text classification, scene recognition, and face analysis, showcases the versatility of CARIn across various model architectures, such as Convolutional Neural Networks and Transformers, and realistic use cases. We observe a substantial enhancement in the fair treatment of the problem's objectives, reaching 1.92x when compared to single-model designs and up to 10.69x in contrast to the state-of-the-art OODIn framework. Additionally, we achieve a significant gain of up to 4.06x over hardware-unaware designs in multi-DNN applications. Finally, our framework sustains its performance while effectively eliminating the time overhead associated with identifying the optimal design in response to environmental challenges.

Horticulture Luminescent Solar Concentrators (HLSCs) represent an innovative concept developed in recent years to promote crop yields, building upon the foundation of traditional Luminescent Solar Concentrators (LSCs). HLSCs are characterized by two distinct properties: spectral conversion and light extraction. Unlike traditional LSCs, HLSCs focus on converting energy from one part of the solar spectrum (typically green) to a specific range (usually red) and aim for the converted photons to exit the device from the bottom surface rather than the edge surfaces. In this study, we start by examining the specific requirements of horticulture to clarify the motivation for using HLSCs. We re-evaluate and propose new optical metrics tailored to HLSCs. Additionally, we analyse potential loss channels for direct red emission and converted red emission. Utilizing Monte Carlo ray tracing method and experimental data, we further explore the factors influencing these loss channels. Our work provides a fundamental discussion on HLSCs and offers design guidelines for future HLSC research.

We present a full one-loop calculation of the gravitino thermal production rate, beyond the so-called hard thermal loop approximation, using the corresponding thermal spectral functions in numerical form on both sides of the light cone. This framework requires a full numerical evaluation. We interpret our results within the framework of a general supergravity-based model, remaining agnostic about the specifics of supersymmetry breaking. In this context, assuming that gravitinos constitute the entirety of the dark matter in the Universe imposes strict constraints on the reheating temperature. For example, with a gluino mass at the current LHC limit, a maximum reheating temperature of $T_\mathrm{reh} \simeq 10^9$ GeV is compatible with a gravitino mass of $m_{3/2} \simeq 1$ TeV. Additionally, with a reheating temperature an order of magnitude lower at $T_\mathrm{reh} \simeq 10^8$ GeV, the common gaugino mass $M_{1/2}$ can range from $2$ to $4 $ TeV within the same gravitino mass range. For much higher values of $M_{1/2}$, which are favored by current accelerator and cosmological data in the context of supersymmetric models, such as $M_{1/2} = 10$ TeV, and for $m_{3/2} \simeq 1$ TeV the reheating temperature compatible with the gravitino dark matter scenario is $ 10^7$ GeV. If other dark matter particles are considered, the reheating temperature could be much lower., Comment: 53 pages, 15 figures

Time reversal (TR) is a promising technique that exploits multipaths for achieving energy focusing in high-frequency wideband communications. In this letter, we focus on a TR scheme facilitated by a reconfigurable intelligent surface (RIS) which, due to the higher frequency and large array aperture, operates in the near-field region. The proposed scheme enriches the propagation environment for the TR in such weak scattering conditions and does not need channel knowledge for the RIS configuration. Specifically, the RIS is employed to create multiple virtual propagation paths that are required to efficiently apply the TR. We derive a performance bound for the proposed scheme under near-field modeling through the received signal-to-noise ratio (SNR) and we examine how various system design parameters affect the performance. We observe that a linear RIS topology maximizes the number of resolvable paths. It is also demonstrated that the proposed scheme improves the SNR, while for a large number of elements it can outperform the conventional passive beamforming at the RIS., Comment: IEEE Communications Letters (5 pages, 4 figures)

Variational Autoencoders (VAEs) are a popular framework for unsupervised learning and data generation. A plethora of methods have been proposed focusing on improving VAEs, with the incorporation of adversarial objectives and the integration of prior learning mechanisms being prominent directions. When it comes to the former, an indicative instance is the recently introduced family of Introspective VAEs aiming at ensuring that a low likelihood is assigned to unrealistic samples. In this study, we focus on the Soft-IntroVAE (S-IntroVAE) and investigate the implication of incorporating a multimodal and learnable prior into this framework. Namely, we formulate the prior as a third player and show that when trained in cooperation with the decoder constitutes an effective way for prior learning, which shares the Nash Equilibrium with the vanilla S-IntroVAE. Furthermore, based on a modified formulation of the optimal ELBO in S-IntroVAE, we develop theoretically motivated regularizations, that is (i) adaptive variance clipping to stabilize training when learning the prior and (ii) responsibility regularization to discourage the formation of inactive prior mode. Finally, we perform a series of targeted experiments on a 2D density estimation benchmark and in an image generation setting comprised of the (F)-MNIST and CIFAR-10 datasets demonstrating the benefit of prior learning in S-IntroVAE in generation and representation learning.

In a changing climate, sustainable agriculture is essential for food security and environmental health. However, it is challenging to understand the complex interactions among its biophysical, social, and economic components. Predictive machine learning (ML), with its capacity to learn from data, is leveraged in sustainable agriculture for applications like yield prediction and weather forecasting. Nevertheless, it cannot explain causal mechanisms and remains descriptive rather than prescriptive. To address this gap, we propose causal ML, which merges ML's data processing with causality's ability to reason about change. This facilitates quantifying intervention impacts for evidence-based decision-making and enhances predictive model robustness. We showcase causal ML through eight diverse applications that benefit stakeholders across the agri-food chain, including farmers, policymakers, and researchers.

This paper investigates the problem of transmit waveform design in the context of a chaotic signal-based self-sustainable reconfigurable intelligent surface (RIS)-aided system for simultaneous wireless information and power transfer (SWIPT). Specifically, we propose a differential chaos shift keying (DCSK)-based RIS-aided point-to-point set-up, where the RIS is partitioned into two non-overlapping surfaces. The elements of the first sub-surface perform energy harvesting (EH), which in turn, provide the required power to the other sub-surface operating in the information transfer (IT) mode. In this framework, by considering a generalized frequency-selective Nakagami-m fading scenario as well as the nonlinearities of the EH process, we derive closed-form analytical expressions for both the bit error rate (BER) at the receiver and the harvested power at the RIS. Our analysis demonstrates, that both these performance metrics depend on the parameters of the wireless channel, the transmit waveform design, and the number of reflecting elements at the RIS, which switch between the IT and EH modes, depending on the application requirements. Moreover, we show that, having more reflecting elements in the IT mode is not always beneficial and also, for a given acceptable BER, we derive a lower bound on the number of RIS elements that need to be operated in the EH mode. Furthermore, for a fixed RIS configuration, we investigate a trade-off between the achievable BER and the harvested power at the RIS and accordingly, we propose appropriate transmit waveform designs. Finally, our numerical results illustrate the importance of our intelligent DCSK-based waveform design on the considered framework., Comment: Submitted for publication in an IEEE journal

MIMO (multiple input, multiple output) approaches are a recent trend in neural network architectures for video restoration problems, where each network evaluation produces multiple output frames. The video is split into non-overlapping stacks of frames that are processed independently, resulting in a very appealing trade-off between output quality and computational cost. In this work we focus on the low-latency setting by limiting the number of available future frames. We find that MIMO architectures suffer from problems that have received little attention so far, namely (1) the performance drops significantly due to the reduced temporal receptive field, particularly for frames at the borders of the stack, (2) there are strong temporal discontinuities at stack transitions which induce a step-wise motion artifact. We propose two simple solutions to alleviate these problems: recurrence across MIMO stacks to boost the output quality by implicitly increasing the temporal receptive field, and overlapping of the output stacks to smooth the temporal discontinuity at stack transitions. These modifications can be applied to any MIMO architecture. We test them on three state-of-the-art video denoising networks with different computational cost. The proposed contributions result in a new state-of-the-art for low-latency networks, both in terms of reconstruction error and temporal consistency. As an additional contribution, we introduce a new benchmark consisting of drone footage that highlights temporal consistency issues that are not apparent in the standard benchmarks., Comment: See the project web page to download the associated videos

X-ray polarization provides a new way to probe accretion geometry in black hole systems. If the accretion geometry of black holes is similar regardless of mass, we should expect the same to be true of their polarization properties. We compare the polarimetric properties of all non-blazar black holes observed with IXPE. We find that their polarization properties are very similar, particularly in the hard state, where the corona dominates. This tentatively supports the idea that stellar and supermassive black holes share a common coronal geometry., Comment: Accepted in ApJ

Computer vision (CV) datasets often exhibit biases that are perpetuated by deep learning models. While recent efforts aim to mitigate these biases and foster fair representations, they fail in complex real-world scenarios. In particular, existing methods excel in controlled experiments involving benchmarks with single-attribute injected biases, but struggle with multi-attribute biases being present in well-established CV datasets. Here, we introduce BAdd, a simple yet effective method that allows for learning fair representations invariant to the attributes introducing bias by incorporating features representing these attributes into the backbone. BAdd is evaluated on seven benchmarks and exhibits competitive performance, surpassing state-of-the-art methods on both single- and multi-attribute benchmarks. Notably, BAdd achieves +27.5% and +5.5% absolute accuracy improvements on the challenging multi-attribute benchmarks, FB-Biased-MNIST and CelebA, respectively.

Defining script types and establishing classification criteria for medieval handwriting is a central aspect of palaeographical analysis. However, existing typologies often encounter methodological challenges, such as descriptive limitations and subjective criteria. We propose an interpretable deep learning-based approach to morphological script type analysis, which enables systematic and objective analysis and contributes to bridging the gap between qualitative observations and quantitative measurements. More precisely, we adapt a deep instance segmentation method to learn comparable character prototypes, representative of letter morphology, and provide qualitative and quantitative tools for their comparison and analysis. We demonstrate our approach by applying it to the Textualis Formata script type and its two subtypes formalized by A. Derolez: Northern and Southern Textualis, Comment: Accepted at ICDAR 2024 Workshop on Computational Paleography (IWCP, 31 August - Athens, Greece)

The exploitation of the strong light-matter coupling regime and exciton-polariton condensates has emerged as a compelling approach to introduce strong interactions and nonlinearities into numerous photonic applications, ranging from low-threshold topological lasers to ultrafast all-optical logic devices. The use of colloidal semiconductor quantum dots with strong three-dimensional confinement as the active material in these microcavities would be highly advantageous due to their versatile structural and compositional tunability and wet-chemical processability, as well as potentially enhanced, confinement-induced polaritonic interactions. Yet, to date, cavity exciton-polariton condensation has neither been achieved with epitaxial nor with colloidal quantum dots. Here, we demonstrate room-temperature polariton condensation in a thin film of monodisperse, colloidal CsPbBr$_3$ quantum dots placed in a tunable optical resonator with a Gaussian-shaped deformation serving as wavelength-scale potential well for the polaritons. The onset of polariton condensation under pulsed optical excitation is manifested in emission by its characteristic superlinear intensity dependence, reduced linewidth, blueshift, and extended temporal coherence. Our results, based on this highly engineerable class of perovskite materials with unique optical properties, pave the way for the development of polaritonic devices for ultrabright coherent light sources and photonic information processing.

Learning with Noisy labels (LNL) poses a significant challenge for the Machine Learning community. Some of the most widely used approaches that select as clean samples for which the model itself (the in-training model) has high confidence, e.g., `small loss', can suffer from the so called `self-confirmation' bias. This bias arises because the in-training model, is at least partially trained on the noisy labels. Furthermore, in the classification case, an additional challenge arises because some of the label noise is between classes that are visually very similar (`hard noise'). This paper addresses these challenges by proposing a method (\textit{CLIPCleaner}) that leverages CLIP, a powerful Vision-Language (VL) model for constructing a zero-shot classifier for efficient, offline, clean sample selection. This has the advantage that the sample selection is decoupled from the in-training model and that the sample selection is aware of the semantic and visual similarities between the classes due to the way that CLIP is trained. We provide theoretical justifications and empirical evidence to demonstrate the advantages of CLIP for LNL compared to conventional pre-trained models. Compared to current methods that combine iterative sample selection with various techniques, \textit{CLIPCleaner} offers a simple, single-step approach that achieves competitive or superior performance on benchmark datasets. To the best of our knowledge, this is the first time a VL model has been used for sample selection to address the problem of Learning with Noisy Labels (LNL), highlighting their potential in the domain., Comment: Accepted to ACMMM2024

The steady improvement of Diffusion Models for visual synthesis has given rise to many new and interesting use cases of synthetic images but also has raised concerns about their potential abuse, which poses significant societal threats. To address this, fake images need to be detected and attributed to their source model, and given the frequent release of new generators, realistic applications need to consider an Open-Set scenario where some models are unseen at training time. Existing forensic techniques are either limited to Closed-Set settings or to GAN-generated images, relying on fragile frequency-based "fingerprint" features. By contrast, we propose a simple yet effective framework that incorporates features from large pre-trained foundation models to perform Open-Set origin attribution of synthetic images produced by various generative models, including Diffusion Models. We show that our method leads to remarkable attribution performance, even in the low-data regime, exceeding the performance of existing methods and generalizes better on images obtained from a diverse set of architectures. We make the code publicly available at: https://github.com/ciodar/UniversalAttribution., Comment: Accepted at ECCV 2024 TWYN workshop

The tunability of individual coupling amplitudes in photonic lattices is highly desirable for photonic Hamiltonian engineering and for studying topological transitions in situ. In this work, we demonstrate the tunneling control between individual lattice sites patterned inside an optical microcavity. The tuning is achieved by applying a voltage to a liquid crystal microcavity possessing photonic Rashba-Dresselhaus spin-orbit coupling. This type of spin-orbit coupling emerges due to the high birefringence of the liquid crystal material and constitutes an artificial gauge field for photons. The proposed technique can be combined with strong-light matter coupling and non-Hermitian physics already established in liquid crystal microcavities.

In the present paper, we outline and expound the fundamental and novel qualitative-cum-philosophical premises, principles, ideas, concepts, constructions and results that originate from our ongoing research project of applying the conceptual panoply and the technical machinery of Abstract Differential Geometry (ADG) to various persistently outstanding issues in Quantum Gravity (QG) research. At the end of the paper, we discuss the potential philosophical repercussions of two possible future research routes that the main stream of our applications of ADG to QG may bifurcate towards in view of three independent, but overlapping, research papers that are currently under development., Comment: Corrected, updated and slightly expanded Version II of the paper, to replace Version I

Regular black holes have become a popular alternative to the singular mathematical black holes predicted by general relativity as they circumvent mathematical pathologies associated with the singularity while preserving crucial black hole features such as the trapping of light. Based on the assumption that semiclassical gravity is valid in the vicinity of their apparent horizons, we review their thermodynamic properties in static scenarios and demonstrate that 1-loop corrections to the Bekenstein-Hawking entropy, accounting for the entanglement entropy of fields, depend on the regularization scheme used. Taking backreaction effects into consideration, we investigate their full dynamical evolution and we examine the behavior of the null energy condition. We find that it is always violated in the vicinity of the outer horizon while being satisfied in the vicinity of the inner horizon, which implies that the trapped spacetime region, as determined from the behavior of null geodesic congruences, is effectively separated into a null energy condition-violating and a non-violating domain. Our findings suggest that quantum effects are more dominant close to the outer apparent horizon and become more pronounced towards the final stages of the evaporation process., Comment: 11 pages, 3 figures. Contribution to the proceedings of the 17th Marcel Grossmann meeting (7-12 July 2024) to be published by World Scientific. Comments welcome!

This paper introduces a novel approach that combines Proper Orthogonal Decomposition (POD) with Thermodynamics-based Artificial Neural Networks (TANN) to capture the macroscopic behavior of complex inelastic systems and derive macroelements in geomechanics. The methodology leverages POD to extract macroscopic Internal State Variables (ISVs) from microscopic state information, thereby enriching the macroscopic state description used to train an energy potential network within the TANN framework. The thermodynamic consistency provided by TANN, combined with the hierarchical nature of POD, allows for accurate modeling of complex, non-linear material behavior and reliable macroscopic geomechanical systems responses. The effectiveness of this approach is validated through applications of increasing complexity, demonstrating its capability to handle various material behaviors and microstructural topologies. These applications include the homogenization of continuous inelastic representative unit cells (RUCs) and the derivation of a macroelement for a geotechnical system involving a monopile in a clay layer subjected to horizontal loading. The results indicate that the proposed POD-TANN methodology not only achieves high accuracy in reproducing stress-strain responses, but also significantly reduces computational costs, making it a practical tool for the multiscale modeling of heterogeneous inelastic systems, and the efficient derivation of macroelements for complex geomechanical problems., Comment: 36 pages, 14 figures, Submitted to International Journal for Numerical and Analytical Methods in Geomechanics

We study the visual quality judgments of human subjects on digital human avatars (sometimes referred to as "holograms" in the parlance of virtual reality [VR] and augmented reality [AR] systems) that have been subjected to distortions. We also study the ability of video quality models to predict human judgments. As streaming human avatar videos in VR or AR become increasingly common, the need for more advanced human avatar video compression protocols will be required to address the tradeoffs between faithfully transmitting high-quality visual representations while adjusting to changeable bandwidth scenarios. During transmission over the internet, the perceived quality of compressed human avatar videos can be severely impaired by visual artifacts. To optimize trade-offs between perceptual quality and data volume in practical workflows, video quality assessment (VQA) models are essential tools. However, there are very few VQA algorithms developed specifically to analyze human body avatar videos, due, at least in part, to the dearth of appropriate and comprehensive datasets of adequate size. Towards filling this gap, we introduce the LIVE-Meta Rendered Human Avatar VQA Database, which contains 720 human avatar videos processed using 20 different combinations of encoding parameters, labeled by corresponding human perceptual quality judgments that were collected in six degrees of freedom VR headsets. To demonstrate the usefulness of this new and unique video resource, we use it to study and compare the performances of a variety of state-of-the-art Full Reference and No Reference video quality prediction models, including a new model called HoloQA. As a service to the research community, we will be publicly releasing the metadata of the new database at https://live.ece.utexas.edu/research/LIVE-Meta-rendered-human-avatar/index.html., Comment: Data will be made available after the paper is accepted. This paper is a preprint of a work currently under a second round of peer review in IEEE TIP

We propose a novel point cloud U-Net diffusion architecture for 3D generative modeling capable of generating high-quality and diverse 3D shapes while maintaining fast generation times. Our network employs a dual-branch architecture, combining the high-resolution representations of points with the computational efficiency of sparse voxels. Our fastest variant outperforms all non-diffusion generative approaches on unconditional shape generation, the most popular benchmark for evaluating point cloud generative models, while our largest model achieves state-of-the-art results among diffusion methods, with a runtime approximately 70% of the previously state-of-the-art PVD. Beyond unconditional generation, we perform extensive evaluations, including conditional generation on all categories of ShapeNet, demonstrating the scalability of our model to larger datasets, and implicit generation which allows our network to produce high quality point clouds on fewer timesteps, further decreasing the generation time. Finally, we evaluate the architecture's performance in point cloud completion and super-resolution. Our model excels in all tasks, establishing it as a state-of-the-art diffusion U-Net for point cloud generative modeling. The code is publicly available at https://github.com/JohnRomanelis/SPVD.git.

Large Language Models (LLMs) have been found to memorize and recite some of the textual sequences from their training set verbatim, raising broad concerns about privacy and copyright issues when using LLMs. This Textual Sequence Memorization (TSM) phenomenon leads to a high demand to regulate LLM output to prevent it from generating certain memorized text to meet user requirements. However, our empirical study reveals that existing methods for TSM erasure fail to forget massive memorized samples without substantially jeopardizing the model utility. To achieve a better trade-off between the effectiveness of TSM erasure and model utility in LLMs, our paper proposes a new framework based on Entropy Maximization with Selective Optimization (EMSO), where the updated weights are chosen with a novel contrastive gradient metric without any participation of additional model or data. Our analysis shows that training with the entropy maximization loss has a more stable optimization process and better keeps model utility than existing methods. The contrastive gradient metric localizes the most influential weight for TSM erasure by taking both the gradient magnitude and direction into consideration. Extensive experiments across three model scales demonstrate that our method excels in handling large-scale forgetting requests while preserving model ability in language generation and reasoning., Comment: 15 pages, 7 figures

We report the discovery of the most eccentric transiting brown dwarf in the brown dwarf desert, TOI02490b. The brown dwarf desert is the lack of brown dwarfs around main sequence stars within $\sim3$~AU and is thought to be caused by differences in formation mechanisms between a star and planet. To date, only $\sim40$ transiting brown dwarfs have been confirmed. \systemt is a $73.6\pm2.4$ \mjupnospace, $1.00\pm0.02$ \rjup brown dwarf orbiting a $1.004_{-0.022}^{+0.031}$ \msunnospace, $1.105_{-0.012}^{+0.012}$ \rsun sun-like star on a 60.33~d orbit with an eccentricity of $0.77989\pm0.00049$. The discovery was detected within \tess sectors 5 (30 minute cadence) and 32 (2 minute and 20 second cadence). It was then confirmed with 31 radial velocity measurements with \feros by the WINE collaboration and photometric observations with the Next Generation Transit Survey. Stellar modelling of the host star estimates an age of $\sim8$~Gyr, which is supported by estimations from kinematics likely placing the object within the thin disc. However, this is not consistent with model brown dwarf isochrones for the system age suggesting an inflated radius. Only one other transiting brown dwarf with an eccentricity higher than 0.6 is currently known in the brown dwarf desert. Demographic studies of brown dwarfs have suggested such high eccentricity is indicative of stellar formation mechanisms., Comment: Accepted for publication in MNRAS, 18 pages, 14 figures

Time series forecasting is an important task in many fields ranging from supply chain management to weather forecasting. Recently, Transformer neural network architectures have shown promising results in forecasting on common time series benchmark datasets. However, application to supply chain demand forecasting, which can have challenging characteristics such as sparsity and cross-series effects, has been limited. In this work, we explore the application of Transformer-based models to supply chain demand forecasting. In particular, we develop a new Transformer-based forecasting approach using a shared, multi-task per-time series network with an initial component applying attention across time series, to capture interactions and help address sparsity. We provide a case study applying our approach to successfully improve demand prediction for a medical device manufacturing company. To further validate our approach, we also apply it to public demand forecasting datasets as well and demonstrate competitive to superior performance compared to a variety of baseline and state-of-the-art forecast methods across the private and public datasets.

Deep Geothermal Energy, Carbon Capture and Storage, and Hydrogen Storage hold considerable promise for meeting the energy sector's large-scale requirements and reducing CO$_2$ emissions. However, the injection of fluids into the Earth's crust, essential for these activities, can induce or trigger earthquakes. In this paper, we highlight a new approach based on Reinforcement Learning for the control of human-induced seismicity in the highly complex environment of an underground reservoir. This complex system poses significant challenges in the control design due to parameter uncertainties and unmodeled dynamics. We show that the reinforcement learning algorithm can interact efficiently with a robust controller, by choosing the controller parameters in real-time, reducing human-induced seismicity and allowing the consideration of further production objectives, \textit{e.g.}, minimal control power. Simulations are presented for a simplified underground reservoir under various energy demand scenarios, demonstrating the reliability and effectiveness of the proposed control-reinforcement learning approach.

This paper introduces an innovative graphical analysis tool for investigating the dynamics of Memristor Cellular Nonlinear Networks (M-CNNs) featuring 2nd-order processing elements, known as M-CNN cells. In the era of specialized hardware catering to the demands of intelligent autonomous systems, the integration of memristors within Cellular Nonlinear Networks (CNNs) has emerged as a promising paradigm due to their exceptional characteristics. However, the standard Dynamic Route Map (DRM) analysis, applicable to 1st-order systems, fails to address the intricacies of 2nd-order M-CNN cell dynamics, as well the 2nd-order DRM (DRM2) exhibits limitations on the graphical illustration of local dynamical properties of the M-CNN cells, e.g. state derivative's magnitude. To address this limitation, we propose a novel integration of M-CNN cell vector field into the cell's phase portrait, enhancing the analysis efficacy and enabling efficient M-CNN cell design. A comprehensive exploration of M-CNN cell dynamics is presented, showcasing the utility of the proposed graphical tool for various scenarios, including bistable and monostable behavior, and demonstrating its superior ability to reveal subtle variations in cell behavior. Through this work, we offer a refined perspective on the analysis and design of M-CNNs, paving the way for advanced applications in edge computing and specialized hardware., Comment: Presented at the 18th IEEE International Workshop on Cellular Nanoscale Networks and their Applications (CNNA'23) and the 8th Memristor and Memristive Symposium

Preservation theorems provide a direct correspondence between the syntactic structure of first-order sentences and the closure properties of their respective classes of models. A line of work has explored preservation theorems relativised to combinatorially tame classes of sparse structures [Atserias et al., JACM 2006; Atserias et al., SiCOMP 2008; Dawar, JCSS 2010; Dawar and Eleftheriadis, 2024]. In this article we initiate the study of preservation theorems for dense graph classes. In contrast to the sparse setting, we show that extension preservation fails on most natural dense classes of low complexity. Nonetheless, we isolate a technical condition which is sufficient for extension preservation to hold, providing a dense analogue to a result of [Atserias et al., SiCOMP 2008]., Comment: 21 pages

In response to stringent emission reduction targets imposed by the International Maritime Organization (IMO) and the European Green Deal's Fit for 55 legislation package, the maritime industry has shifted its focus towards decarbonization. While significant attention has been placed on vessels exceeding 5,000 gross tons (GT), emissions from coastal and short sea shipping, amounting to approximately 13% of global shipping transportation and 15% within the European Union (EU), have not received adequate consideration. This abstract introduces the Zero Emission Sea Transporter (ZEST) project, designed to address this issue by developing a zero-emissions multi-purpose catamaran for short sea route

We develop an algorithm to approximate the time evolution of a probability measure without explicitly learning an operator that governs the evolution. A particular application of interest is discrete measures $\mu_t^N$ that arise from particle systems. In many such situations, the individual particles move chaotically on short time scales, making it difficult to learn the dynamics of a governing operator, but the bulk distribution $\mu_t^N$ approximates an absolutely continuous measure $\mu_t$ that evolves ``smoothly.'' If $\mu_t$ is known on some time interval, then linearized optimal transport theory provides an Euler-like scheme for approximating the evolution of $\mu_t$ using its ``tangent vector field'' (represented as a time-dependent vector field on $\mathbb R^d$), which can be computed as a limit of optimal transport maps. We propose an analog of this Euler approximation to predict the evolution of the discrete measure $\mu_t^N$ (without knowing $\mu_t$). To approximate the analogous tangent vector field, we use a finite difference over a time step that sits between the two time scales of the system -- long enough for the large-$N$ evolution ($\mu_t$) to emerge but short enough to satisfactorily approximate the derivative object used in the Euler scheme. By allowing the limiting behavior to emerge, the optimal transport maps closely approximate the vector field describing the bulk distribution's smooth evolution instead of the individual particles' more chaotic movements. We demonstrate the efficacy of this approach with two illustrative examples, Gaussian diffusion and a cell chemotaxis model, and show that our method succeeds in predicting the bulk behavior over relatively large steps.

We present a novel experimental setup called miniBLAST, which enables systematic and repeatable laboratory tests of structures subjected to blast loads. The explosive source is based on the discharge of high electrical loads on a thin conductor, producing repeatable blast-type shock waves of controlled intensity. Conducting blast experiments under safe laboratory conditions offers significant advantages over large-scale experiments, which are expensive, require specialized personnel, are limited in number, and face repeatability issues. In this work, we provide a comprehensive description of the setup's design rationale and technical characteristics. Moreover, we place particular emphasis on the installation phases, safety, and metrology. The explosive source is analyzed and the signature of blast shock waves is retrieved. Finally, we present an example of a masonry wall subjected to blast loads of varying intensity. We then report its dynamic response in three dimensions and at different time instances. This new experimental setup offers a cost-effective, safe, and repeatable method to study structural dynamics under blast loads, with results that can be upscaled to real structures. It also aids in evaluating numerical models and lays the groundwork for further investigations into blast effects and mitigation.

Motivated by the swift global transmission of infectious diseases, we present a comprehensive framework for network-based epidemic control. Our aim is to curb epidemics using two different approaches. In the first approach, we introduce an optimization strategy that optimally reduces travel rates. We analyze the convergence of this strategy and show that it hinges on the network structure to minimize infection spread. In the second approach, we expand the classic SIR model by incorporating and optimizing quarantined states to strategically contain the epidemic. We show that this problem reduces to the problem of matrix balancing. We establish a link between optimization constraints and the epidemic's reproduction number, highlighting the relationship between network structure and disease dynamics. We demonstrate that applying augmented primal-dual gradient dynamics to the optimal quarantine problem ensures exponential convergence to the KKT point. We conclude by validating our approaches using simulation studies that leverage public data from counties in the state of Massachusetts.

This working paper presents a comprehensive study on the development and analysis of various electricity market models, focusing on continuous, discrete, and fractional-order approaches. The continuous model captures the ongoing interactions between power producers and consumers using differential equations, providing insights into long-term trends and steady-state behaviors. The discrete model, suitable for analyzing scenarios where market events occur at specific time intervals, incorporates memory effects to account for historical behaviors and decisions, offering a realistic representation of short-term market dynamics. The fractional-order model introduces fractional calculus to capture memory effects and hereditary properties, enhancing the model's realism and predictive capability by reflecting the influence of past states on current market behavior.

In this paper, a Full Duplex (FD) eXtremely Large (XL) Multiple-Input Multiple-Output (MIMO) node equipped with reconfigurable metasurface antennas at its transmission and reception sides is considered, which is optimized for simultaneous multi-user communications and sensing in the near-field regime at THz frequencies. We first present a novel Position Error Bound (PEB) analysis for the spatial parameters of multiple targets in the vicinity of the FD node, via the received backscattered data signals, and devise an optimization framework for its metasurface-based precoder and combiner. Then, we formulate and solve an optimization problem aiming at the downlink sum-rate maximization, while simultaneously ensuring a minimum PEB requirement for targets' localization. Our simulation results for a sub-THz system setup validate the joint near-field communications and sensing capability of the proposed FD XL MIMO scheme with metasurfaces antennas, showcasing the interplay of its various design parameters., Comment: 5 pages, 2 figures, to be presented in SPAWC 2024

We study collaboration patterns of Wikidata, one of the world's largest collaborative knowledge graph communities. Wikidata lacks long-term engagement with a small group of priceless members, 0.8%, to be responsible for 80% of contributions. Therefore, it is essential to investigate their behavioural patterns and find ways to enhance their contributions and participation. Previous studies have highlighted the importance of discussions among contributors in understanding these patterns. To investigate this, we analyzed all the discussions on Wikidata and used a mixed methods approach, including statistical tests, network analysis, and text and graph embedding representations. Our research showed that the interactions between Wikidata editors form a small world network where the content of a post influences the continuity of conversations. We also found that the account age of Wikidata members and their conversations are significant factors in their long-term engagement with the project. Our findings can benefit the Wikidata community by helping them improve their practices to increase contributions and enhance long-term participation.

This work introduces EffiSegNet, a novel segmentation framework leveraging transfer learning with a pre-trained Convolutional Neural Network (CNN) classifier as its backbone. Deviating from traditional architectures with a symmetric U-shape, EffiSegNet simplifies the decoder and utilizes full-scale feature fusion to minimize computational cost and the number of parameters. We evaluated our model on the gastrointestinal polyp segmentation task using the publicly available Kvasir-SEG dataset, achieving state-of-the-art results. Specifically, the EffiSegNet-B4 network variant achieved an F1 score of 0.9552, mean Dice (mDice) 0.9483, mean Intersection over Union (mIoU) 0.9056, Precision 0.9679, and Recall 0.9429 with a pre-trained backbone - to the best of our knowledge, the highest reported scores in the literature for this dataset. Additional training from scratch also demonstrated exceptional performance compared to previous work, achieving an F1 score of 0.9286, mDice 0.9207, mIoU 0.8668, Precision 0.9311 and Recall 0.9262. These results underscore the importance of a well-designed encoder in image segmentation networks and the effectiveness of transfer learning approaches., Comment: To be published in IEEE Engineering in Medicine and Biology (EMBC) 2024 conference proceedings

The application of machine learning (ML) in detecting, diagnosing, and treating mental health disorders is garnering increasing attention. Traditionally, research has focused on single modalities, such as text from clinical notes, audio from speech samples, or video of interaction patterns. Recently, multimodal ML, which combines information from multiple modalities, has demonstrated significant promise in offering novel insights into human behavior patterns and recognizing mental health symptoms and risk factors. Despite its potential, multimodal ML in mental health remains an emerging field, facing several complex challenges before practical applications can be effectively developed. This survey provides a comprehensive overview of the data availability and current state-of-the-art multimodal ML applications for mental health. It discusses key challenges that must be addressed to advance the field. The insights from this survey aim to deepen the understanding of the potential and limitations of multimodal ML in mental health, guiding future research and development in this evolving domain.

This paper demonstrates how to use generative models trained for image synthesis as tools for visual data mining. Our insight is that since contemporary generative models learn an accurate representation of their training data, we can use them to summarize the data by mining for visual patterns. Concretely, we show that after finetuning conditional diffusion models to synthesize images from a specific dataset, we can use these models to define a typicality measure on that dataset. This measure assesses how typical visual elements are for different data labels, such as geographic location, time stamps, semantic labels, or even the presence of a disease. This analysis-by-synthesis approach to data mining has two key advantages. First, it scales much better than traditional correspondence-based approaches since it does not require explicitly comparing all pairs of visual elements. Second, while most previous works on visual data mining focus on a single dataset, our approach works on diverse datasets in terms of content and scale, including a historical car dataset, a historical face dataset, a large worldwide street-view dataset, and an even larger scene dataset. Furthermore, our approach allows for translating visual elements across class labels and analyzing consistent changes., Comment: Project Page: https://diff-mining.github.io/ Accepted in ECCV 2024

Language models have been shown to reproduce underlying biases existing in their training data, which is the majority perspective by default. Proposed solutions aim to capture minority perspectives by either modelling annotator disagreements or grouping annotators based on shared metadata, both of which face significant challenges. We propose a framework that trains models without encoding annotator metadata, extracts latent embeddings informed by annotator behaviour, and creates clusters of similar opinions, that we refer to as voices. Resulting clusters are validated post-hoc via internal and external quantitative metrics, as well a qualitative analysis to identify the type of voice that each cluster represents. Our results demonstrate the strong generalisation capability of our framework, indicated by resulting clusters being adequately robust, while also capturing minority perspectives based on different demographic factors throughout two distinct datasets.

Among binary asteroids, (35107) 1991VH stands out as unique given the likely chaotic rotation within its secondary component. The source of this excited dynamical state is unknown. In this work we demonstrate that a past close encounter with Earth could have provided the necessary perturbation to allow the natural internal dynamics, characterized by spin-orbit coupling, to evolve the system into its current dynamical state. In this hypothesis, the secondary of 1991VH was previously in a classical 1:1 spin-orbit resonance with an orbit period likely between 28-35 hours before being perturbed by an Earth encounter within $\sim80,000$ km. We find if the energy dissipation within the secondary is relatively inefficient, this excited dynamical state could persist to today and produce the observed ground-based measurements. Coupled with the orbital history of 1991VH, we can then place a constraint on the tidal dissipation parameters of the secondary., Comment: Accepted for publication in PSJ

The Deep operator network (DeepONet) is a powerful yet simple neural operator architecture that utilizes two deep neural networks to learn mappings between infinite-dimensional function spaces. This architecture is highly flexible, allowing the evaluation of the solution field at any location within the desired domain. However, it imposes a strict constraint on the input space, requiring all input functions to be discretized at the same locations; this limits its practical applications. In this work, we introduce a Resolution Independent Neural Operator (RINO) that provides a framework to make DeepONet resolution-independent, enabling it to handle input functions that are arbitrarily, but sufficiently finely, discretized. To this end, we propose a dictionary learning algorithm to adaptively learn a set of appropriate continuous basis functions, parameterized as implicit neural representations (INRs), from the input data. These basis functions are then used to project arbitrary input function data as a point cloud onto an embedding space (i.e., a vector space of finite dimensions) with dimensionality equal to the dictionary size, which can be directly used by DeepONet without any architectural changes. In particular, we utilize sinusoidal representation networks (SIRENs) as our trainable INR basis functions. We demonstrate the robustness and applicability of RINO in handling arbitrarily (but sufficiently richly) sampled input functions during both training and inference through several numerical examples.

We review old and recent finite de Finetti theorems in total variation distance and in relative entropy, and we highlight their connections with bounds on the difference between sampling with and without replacement. We also establish two new finite de Finetti theorems for exchangeable random vectors taking values in arbitrary spaces. These bounds are tight, and they are independent of the size and the dimension of the underlying space., Comment: 18 pages, no figures

We present Crab X-ray polarization measurements using IXPE data with a total exposure of 300ks, three times more than the initial 2022 discovery paper. Polarization is detected in three times more pulsar phase bins, revealing an S-shaped $+40^\circ$ polarization angle sweep in the main pulse and ${>}1\sigma$ departures from the OPTIMA optical polarization in both pulses, suggesting different radiation mechanisms or sites for the polarized emission at the two wavebands. Our polarization map of the inner nebula reveals a toroidal magnetic field, as seen in prior IXPE analyses. Along the southern jet, the magnetic field orientation relative to the jet axis changes from perpendicular to parallel and the polarization degree decreases by ${\sim}6\%$. These observations may be explained by kink instabilities along the jet or a collision with a dense, jet-deflecting medium at the tip. Using spectropolarimetric analysis, we find asymmetric polarization in the four quadrants of the inner nebula, as expected for a toroidal field geometry, and a spatial correlation between polarization degree and photon index., Comment: 14 pages, 7 figures, 3 tables. Accepted by The Astrophysical Journal

Hidden Markov models (HMMs) offer a robust and efficient framework for analyzing time series data, modelling both the underlying latent state progression over time and the observation process, conditional on the latent state. However, a critical challenge lies in determining the appropriate number of underlying states, often unknown in practice. In this paper, we employ a Bayesian framework, treating the number of states as a random variable and employing reversible jump Markov chain Monte Carlo to sample from the posterior distributions of all parameters, including the number of states. Additionally, we introduce repulsive priors for the state parameters in HMMs, and hence avoid overfitting issues and promote parsimonious models with dissimilar state components. We perform an extensive simulation study comparing performance of models with independent and repulsive prior distributions on the state parameters, and demonstrate our proposed framework on two ecological case studies: GPS tracking data on muskox in Antarctica and acoustic data on Cape gannets in South Africa. Our results highlight how our framework effectively explores the model space, defined by models with different latent state dimensions, while leading to latent states that are distinguished better and hence are more interpretable, enabling better understanding of complex dynamic systems.

Current optical flow and point-tracking methods rely heavily on synthetic datasets. Event cameras are novel vision sensors with advantages in challenging visual conditions, but state-of-the-art frame-based methods cannot be easily adapted to event data due to the limitations of current event simulators. We introduce a novel self-supervised loss combining the Contrast Maximization framework with a non-linear motion prior in the form of pixel-level trajectories and propose an efficient solution to solve the high-dimensional assignment problem between non-linear trajectories and events. Their effectiveness is demonstrated in two scenarios: In dense continuous-time motion estimation, our method improves the zero-shot performance of a synthetically trained model on the real-world dataset EVIMO2 by 29%. In optical flow estimation, our method elevates a simple UNet to achieve state-of-the-art performance among self-supervised methods on the DSEC optical flow benchmark. Our code is available at https://github.com/tub-rip/MotionPriorCMax., Comment: 24 pages, 8 figures, 8 tables, Project Page: https://github.com/tub-rip/MotionPriorCMax

In this paper we describe the geodesics on the K\"ahler cone of the Heisenberg group. Furthermore we also prove that this is not a complete manifold.