Search

Two-dimensional (2D) transition metal dichalcogenides (TMDs) have recently been shown to demonstrate non-volatile resistive switching (NVRS), offering significant advantages such as high-density integration and low energy consumption due to their atomic-scale thinness. In this study, we focus on the adsorption and desorption of metal adatoms, which can modulate the electrical resistivity by several orders of magnitude. We develop material-based relationships of the adsorption energy with electronic and atomic structure descriptors by examining the effects of various transition-metal adsorbates on the surface of TMDs. Our results reveal that adsorption energies of transition metals exhibit consistent trends across different TMDs (MoS$_2$, MoSe$_2$, WS$_2$, WSe$_2$) and can be explained using simple descriptors of the atomic and electronic structure. We propose several models to describe this adsorption process, providing a deeper understanding of a crucial step in the resistive switching mechanism based on formation and dissolution of point defects. Finally, we connect our computed adsorption energies to the switching energy. These findings will help guide rational materials selection for the development of NVRS devices using 2D TMDs.

Subject to the laws of classical physics - the science that governs the design of today's wireless communication systems - there is no need to extract power from a receiver antenna in order to effect communications. If we dispense with a transmission line and, instead, make the front-end electronics colocated with the antenna, then a high input-impedance preamplifier can measure the open-circuit voltage directly on the antenna port without drawing either current or power. Neither Friis' concept of noise figure, nor Shannon information theory, nor electronics technology dictates that we must extract power from an antenna., Comment: 10 pages

Hard X-ray-selected samples of Active Galactic Nuclei (AGN) provide one of the cleanest views of supermassive black hole accretion, but are biased against objects obscured by Compton-thick gas column densities of $N_{\rm H}$ $>$ 10$^{24}$ cm$^{-2}$. To tackle this issue, we present the NuSTAR Local AGN $N_{\rm H}$ Distribution Survey (NuLANDS)$-$a legacy sample of 122 nearby ($z$ $<$ 0.044) AGN primarily selected to have warm infrared colors from IRAS between 25$-$60 $\mu$m. We show that optically classified type 1 and 2 AGN in NuLANDS are indistinguishable in terms of optical [OIII] line flux and mid-to-far infrared AGN continuum bolometric indicators, as expected from an isotropically selected AGN sample, while type 2 AGN are deficient in terms of their observed hard X-ray flux. By testing many X-ray spectroscopic models, we show the measured line-of-sight column density varies on average by $\sim$ 1.4 orders of magnitude depending on the obscurer geometry. To circumvent such issues we propagate the uncertainties per source into the parent column density distribution, finding a directly measured Compton-thick fraction of 35 $\pm$ 9%. By construction, our sample will miss sources affected by severe narrow-line reddening, and thus segregates sources dominated by small-scale nuclear obscuration from large-scale host-galaxy obscuration. This bias implies an even higher intrinsic obscured AGN fraction may be possible, although tests for additional biases arising from our infrared selection find no strong effects on the measured column-density distribution. NuLANDS thus holds potential as an optimized sample for future follow-up with current and next-generation instruments aiming to study the local AGN population in an isotropic manner., Comment: Accepted for publication in ApJ. 50 pages (78 including appendix and bibliography), 21 figures

Research on non-verbal behavior generation for social interactive agents focuses mainly on the believability and synchronization of non-verbal cues with speech. However, existing models, predominantly based on deep learning architectures, often perpetuate biases inherent in the training data. This raises ethical concerns, depending on the intended application of these agents. This paper addresses these issues by first examining the influence of gender on facial non-verbal behaviors. We concentrate on gaze, head movements, and facial expressions. We introduce a classifier capable of discerning the gender of a speaker from their non-verbal cues. This classifier achieves high accuracy on both real behavior data, extracted using state-of-the-art tools, and synthetic data, generated from a model developed in previous work.Building upon this work, we present a new model, FairGenderGen, which integrates a gender discriminator and a gradient reversal layer into our previous behavior generation model. This new model generates facial non-verbal behaviors from speech features, mitigating gender sensitivity in the generated behaviors. Our experiments demonstrate that the classifier, developed in the initial phase, is no longer effective in distinguishing the gender of the speaker from the generated non-verbal behaviors.

Periodic collisions between a star on an inclined orbit around a supermassive black hole and its accretion disk offers a promising explanation for X-ray "quasi-periodic eruptions" (QPEs). Each passage through the disk shocks and compresses gas ahead of the star, which subsequently re-expands above the disk as a quasi-spherical cloud. We present spherically symmetric Monte Carlo radiation transport simulations which follow the production of photons behind the radiation-mediated shock, Comptonization by hot electrons, and the eventual escape of the radiation through the expanding debris. Such one-dimension calculations are approximately justified for thin disks, through which the star of radius $R_{\star}$ passes faster than the shocked gas can flow around the star. For collision speeds $v_{\rm coll} \gtrsim 0.15 c$ and disk surface densities $\Sigma \sim 10^{3}$ g cm$^{-2}$ characteristic of those encountered by stellar orbits consistent with QPE recurrence times, the predicted transient light curves exhibit peak luminosities $\gtrsim 10^{42}$ erg s$^{-1}$ and Comptonized quasi-thermal (Wien-like) spectra which peak at energies $h\nu \sim 100$ eV, broadly consistent with QPE properties. For these conditions, gas and radiation are out of equilibrium and the emission temperature is harder than the blackbody value due to inefficient photon production behind the shock. Alternatively, for higher disk densities and/or lower shock velocities, QPE emission could instead represent the comparatively brief phase shortly after shock break-out, though in this case the bulk of the radiation is thermalized and occurs in the ultraviolet instead of the X-ray band. In either scenario, reproducing the observed eruption properties (duration, luminosity, temperature) requires a large radius $R_{\star} \gtrsim 10R_{\odot}$, which may point to inflation of the star's atmosphere from repeated collisions., Comment: 34 pages, 20 figures. Submitted to ApJ

We present a proof-of-concept of a model comparison approach for analyzing spatio-temporal observations of interacting populations. Our model variants are a collection of structurally similar Bayesian networks. Their distinct Noisy-Or conditional probability distributions describe interactions within the population, with each distribution corresponding to a specific mechanism of interaction. To determine which distributions most accurately represent the underlying mechanisms, we examine the accuracy of each Bayesian network with respect to observational data. We implement such a system for observations of bacterial populations engaged in conjugation, a type of horizontal gene transfer that allows microbes to share genetic material with nearby cells through physical contact. Evaluating cell-specific factors that affect conjugation is generally difficult because of the stochastic nature of the process. Our approach provides a new method for gaining insight into this process. We compare eight model variations for each of three experimental trials and rank them using two different metrics, Comment: 17 pages, 4 figures, 3 tables

The Time-Domain And MultiMessenger (TDAMM) Communications Science Analysis Group (TDAMMCommSAG) was formulated to describe the unique technical challenges of communicating rapidly to and from NASA astrophysics missions studying the most variable, transient, and extreme objects in the Universe. This report describes the study of if and how the transition from current NASA-operated space and ground relays to commercial services will adequately serve these missions. Depending on the individual mission requirements and Concept of Operations (ConOps), TDAMM missions may utilize a rapid low-rate demand access service, a low-rate continuous contact service, low-latency downlink upon demand, or a higher-latency but regular relay service. The specific implementations can vary via space relay or direct to Earth, but requires flexibility and adaptability using modern software infrastructure. The study team reviewed the current state of NASA communications services and future commercial and NASA communications services under study and in development. We explored the communications capabilities driving from the behavior of the astrophysical objects themselves., Comment: 20 page, 1 figure

Precipitation of cold gas due to thermal instability in both galaxy clusters and the circumgalactic medium may regulate AGN feedback. We investigate thermal instability in idealized simulations of the circumgalactic medium with a parameter study of over 600 three-dimensional hydrodynamic simulations of stratified turbulence with cooling, each evolved for 10 Gyr. The entropy profiles are maintained in a steady state via an idealized `thermostat' process, consistent with galaxy cluster entropy profiles. In the presence of external turbulent driving, we find cold gas precipitates, with a strong dependence whether the turbulent driving mechanism is solenoidal, compressive, or purely vertical. In the purely-vertical turbulent driving regime, we find that significant cold gas may form when the cooling time to free-fall time $t_{\rm cool} / t_{\text{ff}} \lesssim 5$. Our simulations with a ratio of $t_{\rm cool} / t_{\text{ff}} \sim 10$ do not precipitate under any circumstances, perhaps because the thermostat mechanism we use maintains a significant non-zero entropy gradient., Comment: 13 pages, 9 figures. Submitted to MNRAS. Simulation movies are available at this URL: https://benwibking.github.io/precipitation.html

We investigate the Lyman-$\alpha$ (Ly$\alpha$) and Lyman continuum (LyC) properties of the Sunburst Arc, a $z=2.37$ gravitationally lensed galaxy with a multiply-imaged, compact region leaking LyC and a triple-peaked Ly$\alpha$ profile indicating direct Ly$\alpha$ escape. Non-LyC-leaking regions show a redshifted Ly$\alpha$ peak, a redshifted and central Ly$\alpha$ peak, or a triple-peaked Ly$\alpha$ profile. We measure the properties of the Ly$\alpha$ profile from different regions of the galaxy using $R\sim5000$ Magellan/MagE spectra. We compare the Ly$\alpha$ spectral properties to LyC and narrowband Ly$\alpha$ maps from Hubble Space Telescope (HST) imaging to explore the subgalactic Ly$\alpha-$LyC connection. We find strong correlations (Pearson correlation coefficient $r>0.6$) between the LyC escape fraction ($f_{\rm esc}^{\rm LyC}$) and Ly$\alpha$ (1) peak separation $v_{\rm{sep}}$, (2) ratio of the minimum flux density between the redshifted and blueshifted Ly$\alpha$ peaks to continuum flux density $f_{\rm{min}}/f_{\rm{cont}}$, and (3) equivalent width. We favor a complex \ion{H}{1} geometry to explain the Ly$\alpha$ profiles from non-LyC-leaking regions and suggest two \ion{H}{1} geometries that could diffuse and/or rescatter the central Ly$\alpha$ peak from the LyC-leaking region into our sightline across transverse distances of several hundred parsecs. Our results emphasize the complexity of Ly$\alpha$ radiative transfer and its sensitivity to the anisotropies of \ion{H}{1} gas on subgalactic scales. Large differences in the physical scales on which we observe spatially variable direct escape Ly$\alpha$, blueshifted Ly$\alpha$, and escaping LyC photons in the Sunburst Arc underscore the importance of resolving the physical scales that govern Ly$\alpha$ and LyC escape., Comment: Submitted to The Astrophysical Journal with revisions from the first referee report. Comments welcome

Quantization has established itself as the primary approach for decreasing the computational and storage expenses associated with Large Language Models (LLMs) inference. The majority of current research emphasizes quantizing weights and activations to enable low-bit general-matrix-multiply (GEMM) operations, with the remaining non-linear operations executed at higher precision. In our study, we discovered that following the application of these techniques, the primary bottleneck in LLMs inference lies in the softmax layer. The softmax operation comprises three phases: exponent calculation, accumulation, and normalization, Our work focuses on optimizing the first two phases. We propose an analytical approach to determine the optimal clipping value for the input to the softmax function, enabling sub-4-bit quantization for LLMs inference. This method accelerates the calculations of both $e^x$ and $\sum(e^x)$ with minimal to no accuracy degradation. For example, in LLaMA1-30B, we achieve baseline performance with 2-bit quantization on the well-known "Physical Interaction: Question Answering" (PIQA) dataset evaluation. This ultra-low bit quantization allows, for the first time, an acceleration of approximately 4x in the accumulation phase. The combination of accelerating both $e^x$ and $\sum(e^x)$ results in a 36.9% acceleration in the softmax operation.

We developed a new class of soft locomotive robots that can self-assemble into a preprogrammed configuration and vary their stiffness afterward in a highly integrated, compact body using contracting-cord particle jamming (CCPJ). We demonstrate this with a tripod-shaped robot, TripodBot, consisting of three CCPJ-based legs attached to a central body. TripodBot is intrinsically soft and can be stored and transported in a compact configuration. On site, it can self-deploy and crawl in a slip-stick manner through the shape morphing of its legs; a simplified analytical model accurately captures the speed. The robot's adaptability is demonstrated by its ability to navigate tunnels as narrow as 61 percent of its deployed body width and ceilings as low as 31 percent of its freestanding height. Additionally, it can climb slopes up to 15 degrees, carry a load of 5 grams (2.4 times its weight), and bear a load 9429 times its weight., Comment: 15 figures

Motivated by recent developments in Hamiltonian variational principles, Hamiltonian variational integrators, and their applications such as to optimization and control, we present a new Type II variational approach for Hamiltonian systems, based on a virtual work principle that enforces the Type II boundary conditions through a combination of essential and natural boundary conditions; particularly, this approach allows us to define this variational principle intrinsically on manifolds. We first develop this variational principle on vector spaces and subsequently extend it to parallelizable manifolds, general manifolds, as well as to the infinite-dimensional setting. Furthermore, we provide a review of variational principles for Hamiltonian systems in various settings as well as their applications.

The Suborbital Imaging Spectrograph for Transition region Irradiance from Nearby Exoplanet host stars (SISTINE) is a rocket-borne ultraviolet (UV) imaging spectrograph designed to probe the radiation environment of nearby stars. SISTINE operates over a bandpass of 98 -- 127 and 130 -- 158 nm, capturing a broad suite of emission lines tracing the full 10$^4$ -- 10$^5$ K formation temperature range critical for reconstructing the full UV radiation field incident on planets orbiting solar-type stars. SISTINE serves as a platform for key technology developments for future ultraviolet observatories. SISTINE operates at moderate resolving power ($R\sim$1500) while providing spectral imaging over an angular extent of $\sim$6', with $\sim$2" resolution at the slit center. The instrument is composed of an f/14 Cassegrain telescope that feeds a 2.1x magnifying spectrograph, utilizing a blazed holographically ruled diffraction grating and a powered fold mirror. Spectra are captured on a large format microchannel plate (MCP) detector consisting of two 113 x 42 mm segments each read out by a cross delay-line anode. Several novel technologies are employed in SISTINE to advance their technical maturity in support of future NASA UV/optical astronomy missions. These include enhanced aluminum lithium fluoride coatings (eLiF), atomic layer deposition (ALD) protective optical coatings, and ALD processed large format MCPs. SISTINE was launched a total of three times with two of the three launches successfully observing targets Procyon A and $\alpha$ Centauri A and B.

Collapsars - rapidly rotating stellar cores that form black holes (BHs) - can power gamma-ray bursts (GRBs) and are proposed to be key contributors to the production of heavy elements in the Universe via the rapid neutron capture process ($r$-process). Previous neutrino-transport collapsar simulations have been unable to unbind neutron-rich material from the disk. However, these simulations have not included magnetic fields or the BH, both of which are essential for launching mass outflows. We present $\nu$H-AMR, a novel neutrino-transport general relativistic magnetohydrodynamic ($\nu$GRMHD) code, which we use to perform the first 3D $\nu$GRMHD collapsar simulations. We find a self-consistent formation of a disk with initially weak magnetic flux, resulting in a low accretion speed and leaving sufficient time for the disk to neutronize. However, once substantial magnetic flux accumulates near the BH, it becomes dynamically important, leading to a magnetically arrested disk that unbinds some of the neutron-rich material. The strong flux also accelerates the accretion speed, preventing further disk neutronization. The neutron-rich disk ejecta collides with the infalling stellar gas, generating a shocked cocoon with an electron fraction, $Y_\text{e}\gtrsim0.2$. Continuous mixing between the cocoon and neutron-poor stellar gas incrementally raises the outflow $Y_\text{e}$, but the final $r$-process yield is determined earlier at the point of neutron capture freeze-out. Our models require extreme magnetic fluxes and mass accretion rates to eject neutron-rich material ($Y_\text{e}\lesssim0.3$), implying very high $r$-process ejecta masses $M_\text{ej}\lesssim{}M_\odot$. Future work will explore under what conditions more typical collapsar engines become $r$-process factories., Comment: 13 pages, 5 figures, 1 table (plus 7 pages and 7 figures in Appendix) uses aastex62.cls. Submitted to ApJ, comments welcome. Movies available at https://www.youtube.com/playlist?list=PLVsp-mE5pJ_UBsLD1LU1ZZCzPoLgo6wUP

We present a method for gradient computation in quantum algorithms implemented on linear optical quantum computing platforms. While parameter-shift rules have become a staple in qubit gate-based quantum computing for calculating gradients, their direct application to photonic platforms has been hindered by the non-unitary nature of differentiated phase-shift operators in Fock space. We introduce a photonic parameter-shift rule that overcomes this limitation, providing an exact formula for gradient computation in linear optical quantum processors. Our method scales linearly with the number of input photons and utilizes the same parameterized photonic circuit with shifted parameters for each evaluation. This advancement bridges a crucial gap in photonic quantum computing, enabling efficient gradient-based optimization for variational quantum algorithms on near-term photonic quantum processors. We demonstrate the efficacy of our approach through numerical simulations in quantum chemistry and generative modeling tasks, showing superior optimization performance as well as robustness to noise from finite sampling and photon distinguishability compared to other gradient-based and gradient-free methods., Comment: 12 pages (incl. 4 pages for appendices), 3 figures

Language models are widely used in computational psycholinguistics to test theories that relate the negative log probability (the surprisal) of a region of interest (a substring of characters) under a language model to its cognitive cost experienced by readers, as operationalized, for example, by gaze duration on the region. However, the application of modern language models to psycholinguistic studies is complicated by the practice of using tokenization as an intermediate step in training a model. Doing so results in a language model over token strings rather than one over character strings. Vexingly, regions of interest are generally misaligned with these token strings. The paper argues that token-level language models should be (approximately) marginalized into character-level language models before they are used in psycholinguistic studies to compute the surprisal of a region of interest; then, the marginalized character-level language model can be used to compute the surprisal of an arbitrary character substring, which we term a focal area, that the experimenter may wish to use as a predictor. Our proposal of marginalizing a token-level model into a character-level one solves this misalignment issue independently of the tokenization scheme. Empirically, we discover various focal areas whose surprisal is a better psychometric predictor than the surprisal of the region of interest itself., Comment: Main conference long paper at EMNLP 2024

This paper explores the potential of synthetic physical Copy Detection Patterns (CDP) to improve the robustness of anti-counterfeiting systems. By leveraging synthetic physical CDP, we aim at enhancing security and cost-effectiveness across various real-world applications. Our research demonstrates that synthetic CDP offer substantial improvements in authentication accuracy compared to one based on traditional digital templates. We conducted extensive tests using both a scanner and a diverse range of mobile phones, validating our approach through ROC analysis. The results indicate that synthetic CDP can reliably differentiate between original and fake samples, making this approach a viable solution for real-world applications, though requires an additional research to make this technology scalable across a variety of imaging devices.

We have discovered a triply eclipsing triple-star system, TIC 290061484, with the shortest known outer period, Pout, of only 24.5 days. This "eclipses" the previous record set by lambda Tauri at 33.02 days, which held for 68 yr. The inner binary, with an orbital period of Pin = 1.8 days, produces primary and secondary eclipses and exhibits prominent eclipse timing variations with the same periodicity as the outer orbit. The tertiary star eclipses, and is eclipsed by, the inner binary with pronounced asymmetric profiles. The inclinations of both orbits evolve on observable timescales such that the third-body eclipses exhibit dramatic depth variations in TESS data. A photodynamical model provides a complete solution for all orbital and physical parameters of the triple system, showing that the three stars have masses of 6.85, 6.11, and 7.90 MSun, radii near those corresponding to the main sequence, and Teff in the range of 21,000-23,700 K. Remarkably, the model shows that the triple is in fact a subsystem of a hierarchical 2+1+1 quadruple with a distant fourth star. The outermost star has a period of ~3200 days and a mass comparable to the stars in the inner triple. In ~20 Myr, all three components of the triple subsystem will merge, undergo a Type II supernova explosion, and leave a single remnant neutron star. At the time of writing, TIC 290061484 is the most compact triple system and one of the tighter known compact triples (i.e., Pout/Pin = 13.7)., Comment: 27 pages, 13 figures, 6 tables

Despite extensive safety training, LLMs are vulnerable to adversarial inputs. In this work, we introduce a simple but powerful attack paradigm, bijection learning, that yields a practically endless set of jailbreak prompts. We exploit language models' advanced reasoning capabilities to teach them invertible languages (bijections) in context, pass encoded queries to the model to bypass built-in safety mechanisms, and finally decode responses back into English, yielding helpful replies to harmful requests. Our approach proves effective on a wide range of frontier language models and harm categories. Bijection learning is an automated and universal attack that grows stronger with scale: larger models with more advanced reasoning capabilities are more susceptible to bijection learning jailbreaks despite stronger safety mechanisms.

In longitudinal studies using routinely collected data, such as electronic health records (EHRs), patients tend to have more measurements when they are unwell; this informative observation pattern may lead to bias. While semi-parametric approaches to modelling longitudinal data subject to irregular observation are known to be sensitive to misspecification of the visit process, parametric models may provide a more robust alternative. Robustness of parametric models on the outcome alone has been assessed under the assumption that the visit intensity is independent of the time since the last visit, given the covariates and random effects. However, this assumption of a memoryless visit process may not be realistic in the context of EHR data. In a special case which includes memory embedded into the visit process, we derive an expression for the bias in parametric models for the outcome alone and use this to identify factors that lead to increasing bias. Using simulation studies, we show that this bias is often small in practice. We suggest diagnostics for identifying the specific cases when the outcome model may be susceptible to meaningful bias, and propose a novel joint model of the outcome and visit processes that can eliminate or reduce the bias. We apply these diagnostics and the joint model to a study of juvenile dermatomyositis. We recommend that future studies using EHR data avoid relying only on the outcome model and instead first evaluate its appropriateness with our proposed diagnostics, applying our proposed joint model if necessary., Comment: 43 pages, 4 tables, 6 figures

Active shooter incidents in schools cause widespread attention across the nation. Students, faculty, and staff on campuses could be involved with these shootings, as victims, perpetrators, etc.[1]. These gun-related crimes jeopardize school safety. From 1999 to 2024, there have been approximately 43 mass school shootings, with over 500 school shootings altogether. By definition, mass shooting is defined as any event where four or more people are shot with a gun, but not counting the perpetrator. By studying school shooting cases, we concluded that most of the time, the shootings occur inside the classrooms. Existing research that includes statistical analysis usually focuses on public mass shootings or just shooting incidents that have occurred in the past and there are hardly any articles focusing on school mass shootings. This leads to schools being more vulnerable to mass shootings in the future. In this research, we have gathered school shooting data from various resources to analyze the results. By interpreting these data and conducting various statistical analysis, this will ultimately help the law enforcement to better prepare for future school shootings., Comment: 6 pages, 6 figures, conference

The development of large-scale quantum networks requires reliable quantum channels, the quality of which can be quantified by the framework of quantum process tomography. In this work, we leverage ancilla-assisted process tomography and Bayesian inference to probe a 1.6 km deployed fiber-optic link. We send one of two polarization-entangled photons from Alice in one building to Bob in another, exploiting the local qubit as an ancilla system to characterize the corresponding quantum channel. Monitoring over a 24 h period returns a steady process fidelity of 95.1(1)%, while controllable spectral filtering with passbands from 0.025-4.38 THz finds fidelities that first increase, then level off with bandwidth, suggesting both stable operation with time and minimal polarization mode dispersion. To our knowledge, these results represent the first AAPT of a deployed quantum link, revealing a valuable tool for in situ analysis of entanglement-based quantum networks.

Transfer learning is a common practice that alleviates the need for extensive data to train neural networks. It is performed by pre-training a model using a source dataset and fine-tuning it for a target task. However, not every source dataset is appropriate for each target dataset, especially for time series. In this paper, we propose a novel method of selecting and using multiple datasets for transfer learning for time series classification. Specifically, our method combines multiple datasets as one source dataset for pre-training neural networks. Furthermore, for selecting multiple sources, our method measures the transferability of datasets based on shapelet discovery for effective source selection. While traditional transferability measures require considerable time for pre-training all the possible sources for source selection of each possible architecture, our method can be repeatedly used for every possible architecture with a single simple computation. Using the proposed method, we demonstrate that it is possible to increase the performance of temporal convolutional neural networks (CNN) on time series datasets., Comment: Accepted at International Conference on Pattern Recognition 2024 (ICPR 2024)

Generative AI, in particular large transformer models, are increasingly driving HPC system design in science and industry. We analyze performance characteristics of such transformer models and discuss their sensitivity to the transformer type, parallelization strategy, and HPC system features (accelerators and interconnects). We utilize a performance model that allows us to explore this complex design space and highlight its key components. We find that different transformer types demand different parallelism and system characteristics at different training regimes. Large Language Models are performant with 3D parallelism and amplify network needs only at pre-training scales with reduced dependence on accelerator capacity and bandwidth. On the other hand, long-sequence transformers, representative of scientific foundation models, place a more uniform dependence on network and capacity with necessary 4D parallelism. Our analysis emphasizes the need for closer performance modeling of different transformer types keeping system features in mind and demonstrates a path towards this. Our code is available as open-source., Comment: 17 pages, PMBS 2024

The Charting Cluster Construction with VUDS and ORELSE (C3VO) survey is an ongoing imaging and spectroscopic campaign aiming to map out the growth of structure up to $z\sim5$ and was born from the combination of the VIMOS Ultra Deep Survey (VUDS) and the Observations of Redshift Evolution in Large-Scale Environments (ORELSE) survey. As we previously accomplished with the ORELSE survey, we apply our technique known as Voronoi tessellation Monte-Carlo (VMC) mapping to search for serendipitous galaxy overdensities at $2 14$. In the C3VO fields, we were able to recover many of the previously known structures in the literature as well as find hundreds of new overdensity candidates, once again demonstrating the powerful capabilities of VMC mapping when applied to wide-field optical and infrared galaxy evolution surveys at ever higher redshifts., Comment: 36 pages, 16 figures, 5 tables, submitted to ApJ

Given an input painting, we reconstruct a time-lapse video of how it may have been painted. We formulate this as an autoregressive image generation problem, in which an initially blank "canvas" is iteratively updated. The model learns from real artists by training on many painting videos. Our approach incorporates text and region understanding to define a set of painting "instructions" and updates the canvas with a novel diffusion-based renderer. The method extrapolates beyond the limited, acrylic style paintings on which it has been trained, showing plausible results for a wide range of artistic styles and genres., Comment: Project Page: https://inversepainting.github.io

Background: Large language models (LLMs) are trained to follow directions, but this introduces a vulnerability to blindly comply with user requests even if they generate wrong information. In medicine, this could accelerate the generation of misinformation that impacts human well-being. Objectives/Methods: We analyzed compliance to requests to generate misleading content about medications in settings where models know the request is illogical. We investigated whether in-context directions and instruction-tuning of LLMs to prioritize logical reasoning over compliance reduced misinformation risk. Results: While all frontier LLMs complied with misinformation requests, both prompt-based and parameter-based approaches can improve the detection of logic flaws in requests and prevent the dissemination of medical misinformation. Conclusion: Shifting LLMs to prioritize logic over compliance could reduce risks of exploitation for medical misinformation., Comment: Submitted for Review

In complex auditory environments, the human auditory system possesses the remarkable ability to focus on a specific speaker while disregarding others. In this study, a new model named SWIM, a short-window convolution neural network (CNN) integrated with Mamba, is proposed for identifying the locus of auditory attention (left or right) from electroencephalography (EEG) signals without relying on speech envelopes. SWIM consists of two parts. The first is a short-window CNN (SW$_\text{CNN}$), which acts as a short-term EEG feature extractor and achieves a final accuracy of 84.9% in the leave-one-speaker-out setup on the widely used KUL dataset. This improvement is due to the use of an improved CNN structure, data augmentation, multitask training, and model combination. The second part, Mamba, is a sequence model first applied to auditory spatial attention decoding to leverage the long-term dependency from previous SW$_\text{CNN}$ time steps. By joint training SW$_\text{CNN}$ and Mamba, the proposed SWIM structure uses both short-term and long-term information and achieves an accuracy of 86.2%, which reduces the classification errors by a relative 31.0% compared to the previous state-of-the-art result. The source code is available at https://github.com/windowso/SWIM-ASAD., Comment: accepted by SLT 2024

We propose a new accelerated first-order method for convex optimization under non-Euclidean smoothness assumptions. In contrast to standard acceleration techniques, our approach uses primal-dual iterate sequences taken with respect to differing norms, which are then coupled using an implicitly determined interpolation parameter. For $\ell_p$ norm smooth problems in $d$ dimensions, our method provides an iteration complexity improvement of up to $O(d^{1-\frac{2}{p}})$ in terms of calls to a first-order oracle, thereby allowing us to circumvent long-standing barriers in accelerated non-Euclidean steepest descent.

Two-dimensional (2D) 1T-VSe$_2$ has prompted significant interest due to the discrepancies regarding alleged ferromagnetism (FM) at room temperature, charge density wave (CDW) states and the interplay between the two. We employed a combined Diffusion Monte Carlo (DMC) and density functional theory (DFT) approach to accurately investigate the magnetic properties and response of strain of monolayer 1T-VSe$_2$. Our calculations show the delicate competition between various phases, revealing critical insights into the relationship between their energetic and structural properties. We went on to perform Classical Monte Carlo simulations informed by our DMC and DFT results, and found the magnetic transition temperature ($T_c$) of the undistorted (non-CDW) FM phase to be 228 K and the distorted (CDW) phase to be 68 K. Additionally, we studied the response of biaxial strain on the energetic stability and magnetic properties of various phases of 2D 1T-VSe$_2$ and found that small amounts of strain can enhance the $T_c$, suggesting a promising route for engineering and enhancing magnetic behavior. Finally, we synthesized 1T-VSe$_2$ and performed Raman spectroscopy measurements, which were in close agreement with our calculated results. Our work emphasizes the role of highly accurate DMC methods in advancing the understanding of monolayer 1T-VSe$_2$ and provides a robust framework for future studies of 2D magnetic materials.

Future distributed quantum systems and networks are likely to rely, at least in part, on the existing fiber infrastructure for entanglement distribution; hence, a precise understanding of the adverse effects of imperfections in optical fibers on entanglement is essential to their operation. Here, we consider maximally entangled polarization qubits and study the decay of the entanglement caused by spatial fluctuations in the refractive index of optical fibers. We study this entanglement dynamics using the spin-boson model and numerically solve our system of equations using the hierarchical equations of motion (HEOM) formalism. We show that within the range of practically relevant system parameters, our developed model exhibits both Markovian and non-Markovian entanglement decay behavior. Further, to counter the observed entanglement decay, we propose the implementation of dynamical decoupling in optical fibers using spaced half waveplates. In particular, we numerically model the time-dependent Hamiltonians of the Carr-Purcell-Meiboom-Gill and Uhrig dynamical decoupling schemes and show a reduced rate of entanglement decay even with sparsely spaced half waveplates along the length of optical fiber. Finally, we evaluate the performance of these two schemes in multiple system configurations., Comment: 10 pages, 8 figures

Multilingual Automatic Speech Recognition (ASR) models are typically evaluated in a setting where the ground-truth language of the speech utterance is known, however, this is often not the case for most practical settings. Automatic Spoken Language Identification (SLID) models are not perfect and misclassifications have a substantial impact on the final ASR accuracy. In this paper, we present a simple and effective N-best re-ranking approach to improve multilingual ASR accuracy for several prominent acoustic models by employing external features such as language models and text-based language identification models. Our results on FLEURS using the MMS and Whisper models show spoken language identification accuracy improvements of 8.7% and 6.1%, respectively and word error rates which are 3.3% and 2.0% lower on these benchmarks.

A wide range of physical phenomena exhibit auxiliary admissibility criteria, such as conservation of entropy or various energies, which arise implicitly under exact solution of their governing PDEs. However, standard temporal schemes, such as classical Runge-Kutta (RK) methods, do not enforce these constraints, leading to a loss of accuracy and stability. Projection is an efficient way to address this shortcoming by correcting the RK solution at the end of each time step. Here we introduce a novel projection method for explicit RK schemes, called a \textit{quasi-orthogonal} projection method. This method can be employed for systems containing a single (not necessarily convex) invariant functional, for dissipative systems, and for the systems containing multiple invariants. It works by projecting the orthogonal search direction(s) into the solution space spanned by the RK stage derivatives. With this approach linear invariants of the problem are preserved, the time step size remains fixed, additional computational cost is minimal, and these optimal search direction(s) preserve the order of accuracy of the base RK method. This presents significant advantages over existing projection methods. Numerical results demonstrate that these properties are observed in practice for a range of applications.

Merging has become a widespread way to cheaply combine individual models into a single model that inherits their capabilities and attains better performance. This popularity has spurred rapid development of many new merging methods, which are typically validated in disparate experimental settings and frequently differ in the assumptions made about model architecture, data availability, and computational budget. In this work, we characterize the relative merits of different merging methods by evaluating them in a shared experimental setting and precisely identifying the practical requirements of each method. Specifically, our setting focuses on using merging for compositional generalization of capabilities in image classification, image generation, and natural language processing. Additionally, we measure the computational costs of different merging methods as well as how they perform when scaling the number of models being merged. Taken together, our results clarify the state of the field of model merging and provide a comprehensive and rigorous experimental setup to test new methods.

Current efforts to detect nuclear detonations and correctly categorize explosion sources with ground- and space-collected discriminants presents challenges that remain unaddressed by the Event Categorization Matrix (ECM) model. Smaller events (lower yield explosions) often include only sparse observations among few modalities and can therefore lack a complete set of discriminants. The covariance structures can also vary significantly between such observations of event (source-type) categories. Both obstacles are problematic for ``classic'' ECM. Our work addresses this gap and presents a Bayesian update to the previous ECM model, termed B-ECM, which can be trained on partial observations and does not rely on a pooled covariance structure. We further augment ECM with Bayesian Decision Theory so that false negative or false positive rates of an event categorization can be reduced in an intuitive manner. To demonstrate improved categorization rates with B-ECM, we compare an array of B-ECM and classic ECM models with multiple performance metrics that leverage Monte Carlo experiments. We use both synthetic and real data. Our B-ECM models show consistent gains in overall accuracy and a lower false negative rates relative to the classic ECM model. We propose future avenues to improve B-ECM that expand its decision-making and predictive capability., Comment: Main text 20 pages, 5 figures

The vast amounts of digital content captured from the real world or AI-generated media necessitate methods for copyright protection, traceability, or data provenance verification. Digital watermarking serves as a crucial approach to address these challenges. Its evolution spans three generations: handcrafted, autoencoder-based, and foundation model based methods. While the robustness of these systems is well-documented, the security against adversarial attacks remains underexplored. This paper evaluates the security of foundation models' latent space digital watermarking systems that utilize adversarial embedding techniques. A series of experiments investigate the security dimensions under copy and removal attacks, providing empirical insights into these systems' vulnerabilities. All experimental codes and results are available at https://github.com/vkinakh/ssl-watermarking-attacks .

Large Language Models have advanced clinical Natural Language Generation, creating opportunities to manage the volume of medical text. However, the high-stakes nature of medicine requires reliable evaluation, which remains a challenge. In this narrative review, we assess the current evaluation state for clinical summarization tasks and propose future directions to address the resource constraints of expert human evaluation.

We develop formal privacy mechanisms for releasing statistics from data with many outlying values, such as income data. These mechanisms ensure that a per-record differential privacy guarantee degrades slowly in the protected records' influence on the statistics being released. Formal privacy mechanisms generally add randomness, or "noise," to published statistics. If a noisy statistic's distribution changes little with the addition or deletion of a single record in the underlying dataset, an attacker looking at this statistic will find it plausible that any particular record was present or absent, preserving the records' privacy. More influential records -- those whose addition or deletion would change the statistics' distribution more -- typically suffer greater privacy loss. The per-record differential privacy framework quantifies these record-specific privacy guarantees, but existing mechanisms let these guarantees degrade rapidly (linearly or quadratically) with influence. While this may be acceptable in cases with some moderately influential records, it results in unacceptably high privacy losses when records' influence varies widely, as is common in economic data. We develop mechanisms with privacy guarantees that instead degrade as slowly as logarithmically with influence. These mechanisms allow for the accurate, unbiased release of statistics, while providing meaningful protection for highly influential records. As an example, we consider the private release of sums of unbounded establishment data such as payroll, where our mechanisms extend meaningful privacy protection even to very large establishments. We evaluate these mechanisms empirically and demonstrate their utility.

Understanding human mobility patterns has traditionally been a complex challenge in transportation modeling. Due to the difficulties in obtaining high-quality training datasets across diverse locations, conventional activity-based models and learning-based human mobility modeling algorithms are particularly limited by the availability and quality of datasets. Furthermore, current research mainly focuses on the spatial-temporal travel pattern but lacks an understanding of the semantic information between activities, which is crucial for modeling the interdependence between activities. In this paper, we propose an innovative Large Language Model (LLM) empowered human mobility modeling framework. Our proposed approach significantly reduces the reliance on detailed human mobility statistical data, utilizing basic socio-demographic information of individuals to generate their daily mobility patterns. We have validated our results using the NHTS and SCAG-ABM datasets, demonstrating the effective modeling of mobility patterns and the strong adaptability of our framework across various geographic locations.

We investigate the co-evolution of the angular momentum of Milky Way-like galaxies, their circumgalactic gas, and their dark matter halos using zoom-in simulations from the Figuring Out Gas & Galaxies in Enzo (FOGGIE) suite. We examine how the magnitude and orientation of the angular momentum varies over time within the halo and between the components of mass. From z~2 to today, and in general across the simulated halos, the specific angular momenta of the central galaxies and the cool gas in their circumgalactic media (T < 10^5 K) increase together. Over that same period, the specific angular momenta of the hot (>10^6 K) and dark components of the halo change minimally. By z~1, the central galaxies have generally lost association with the angular momentum of their full dark matter halo -- both in magnitude and orientation. We find a wide distribution of angular momentum orientations in the halo, varying by up to 180 degrees over small (~tens of kpc) scales and between the different components of mass. The net angular momenta of the galaxies, their circumgalactic gas, and their dark matter halos are generally misaligned with one another at all cosmic times. The present-day orientation of the central galaxies are established at late times (after z=1), after the rates of cosmic accretion and mergers decline and the disks are able to settle and stabilize their orientation., Comment: 35 pages, 14 figures, submitted to ApJ

The advent of new and near-future observatories probing the earliest epochs of the Universe has opened the opportunity to investigate the formation and growth of the first massive black holes (MBHs). Additionally, the use of high resolution cosmological simulations to investigate these high-redshift environments is needed to predict the dark matter halos in which these MBH seeds will form. We use the $\textit{Renaissance}$ simulations to analyse the formation and growth of so-called heavy seed black holes. Other past work has investigated the formation and growth of light (black hole) seeds with $\textit{Renaissance}$ and found that these black holes do not grow in the environments in which they reside. In this work we seed MBHs, in post-processing, and track accretion onto the MBHs as well as mergers with other MBHs at high-redshift. We show that the heavy seeds struggle to achieve high accretion rates with only the most massive black holes ($\gtrsim 10^5 \text{M}_\odot$) growing at close to the Eddington limit under optimistic conditions. Despite the lack of significant growth for these early MBHs, the signals from their merger events will be sufficiently strong (SNR $\sim 10^2$) to be probed by the next generation of gravitational wave observatories, such as $\textit{LISA}$. We predict that $\textit{LISA}$ will observe of the order of $10$ MBH merger events per year where the mergers occur at z $\gtrsim$ 10 or at least begin their early inspiral phase at z $\gtrsim$ 10., Comment: 15 pages, 10 figures, Submitted to the Open Journal of Astrophysics for review

Though the BB84 protocol has provable security over a noiseless quantum channel, the security is not proven over current noisy technology. The level of tolerable error on such systems is still unclear, as is how much information about a raw key may be obtained by an eavesdropper. We develop a reproducible test to determine the security--or lack thereof--of the protocol in practice. This enables us to obtain an experimental estimate of the information that can be obtained using asymmetric phase-covariant cloning to eavesdrop on the BB84 protocol.

Recently, large language models (LLMs) have demonstrated exceptional capabilities in serving as the foundation for AI assistants. One emerging application of LLMs, navigating through websites and interacting with UI elements across various web pages, remains somewhat underexplored. We introduce Steward, a novel LLM-powered web automation tool designed to serve as a cost-effective, scalable, end-to-end solution for automating web interactions. Traditional browser automation frameworks like Selenium, Puppeteer, and Playwright are not scalable for extensive web interaction tasks, such as studying recommendation algorithms on platforms like YouTube and Twitter. These frameworks require manual coding of interactions, limiting their utility in large-scale or dynamic contexts. Steward addresses these limitations by integrating LLM capabilities with browser automation, allowing for natural language-driven interaction with websites. Steward operates by receiving natural language instructions and reactively planning and executing a sequence of actions on websites, looping until completion, making it a practical tool for developers and researchers to use. It achieves high efficiency, completing actions in 8.52 to 10.14 seconds at a cost of $0.028 per action or an average of $0.18 per task, which is further reduced to 4.8 seconds and $0.022 through a caching mechanism. It runs tasks on real websites with a 40% completion success rate. We discuss various design and implementation challenges, including state representation, action sequence selection, system responsiveness, detecting task completion, and caching implementation.

Recent advances in large language models have enabled the creation of highly effective chatbots, which may serve as a platform for targeted advertising. This paper investigates the risks of personalizing advertising in chatbots to their users. We developed a chatbot that embeds personalized product advertisements within LLM responses, inspired by similar forays by AI companies. Our benchmarks show that ad injection impacted certain LLM attribute performance, particularly response desirability. We conducted a between-subjects experiment with 179 participants using chabots with no ads, unlabeled targeted ads, and labeled targeted ads. Results revealed that participants struggled to detect chatbot ads and unlabeled advertising chatbot responses were rated higher. Yet, once disclosed, participants found the use of ads embedded in LLM responses to be manipulative, less trustworthy, and intrusive. Participants tried changing their privacy settings via chat interface rather than the disclosure. Our findings highlight ethical issues with integrating advertising into chatbot responses

Electron microscopy prevalently uses energy-dispersive x-ray spectroscopy (EDS) and electron energy loss spectroscopy (EELS) for elemental analysis. EDS and EELS energy resolutions are commonly between 30-100 eV or 0.01-1 eV, respectively. Large solid angle EDS detector technology has increased collection efficiency to enable precision spectroscopy via averaging of 0.02-0.1 eV. This improved precision gives access to chemical shifts; examples are shown in compounds of Al, Ti, and W. EDS can now detect chemical information in a complementary parameter space (accelerating voltage, thickness, atomic number) to that covered by EELS., Comment: 6 main text pages, 18 supplementary text pages

Modern vehicles have become sophisticated computation and sensor systems, as evidenced by advanced driver assistance systems, in-car infotainment, and autonomous driving capabilities. They collect and process vast amounts of data through various embedded subsystems. One significant player in this landscape is Android Automotive OS (AAOS), which has been integrated into over 100M vehicles and has become a dominant force in the in-vehicle infotainment market. With this extensive data collection, privacy has become increasingly crucial. The volume of data gathered by these systems raises questions about how this information is stored, used, and protected, making privacy a critical issue for manufacturers and consumers. However, very little has been done on vehicle data privacy. This paper focuses on the privacy implications of AAOS, examining the exact nature and scope of data collection and the corresponding privacy policies from the original equipment manufacturers (OEMs). We develop a novel automotive privacy analysis tool called PriDrive which employs three methodological approaches: network traffic inspection, and both static and dynamic analyses of Android images using rooted emulators from various OEMs. These methodologies are followed by an assessment of whether the collected data types were properly disclosed in OEMs and 3rd party apps' privacy policies (to identify any discrepancies or violations). Our evaluation on three different OEM platforms reveals that vehicle speed is collected at a sampling rate of roughly 25 Hz. Other properties such as model info, climate & AC, and seat data are collected in a batch 30 seconds into vehicle startup. In addition, several vehicle property types were collected without disclosure in their respective privacy policies. For example, OEM A's policies only covers 110 vehicle properties or 13.02% of the properties found in our static analysis.

Ono, Saad, and Saikia introduced moments for hypergeometric functions over finite fields. These authors and Pujahari have established moments for several $_{2}F_{1}$ and $_{3}F_{2}$ cases. We study $_{2}F_{1}$ and $_{3}F_{2}$ cases when the hypergeometric data is defined over $\mathbb{Q}$ and primitive. Moments are established in several cases using Hecke trace formulas of hypergeometric origin recently shown by Hoffman, Li, Long, and Tu. We also obtain several algebraic formulas in the finite-field setting and present conjectures for additional $_{2}F_{1}$ and $_{3}F_{2}$ moments.

The growing prevalence of inverter-based resources (IBRs) for renewable energy integration and electrification greatly challenges power system dynamic analysis. To account for both synchronous generators (SGs) and IBRs, this work presents an approach for learning the model of an individual dynamic component. The recurrent neural network (RNN) model is used to match the recursive structure in predicting the key dynamical states of a component from its terminal bus voltage and set-point input. To deal with the fast transients especially due to IBRs, we develop a Stable Integral (SI-)RNN to mimic high-order integral methods that can enhance the stability and accuracy for the dynamic learning task. We demonstrate that the proposed SI-RNN model not only can successfully predict the component's dynamic behaviors, but also offers the possibility of efficiently computing the dynamic sensitivity relative to a set-point change. These capabilities have been numerically validated based on full-order Electromagnetic Transient (EMT) simulations on a small test system with both SGs and IBRs, particularly for predicting the dynamics of grid-forming inverters.

Background: Human skin texture has yet to be quantified for diagnostic purposes. Here, the surface metrology of seborrheic keratoses is investigated with an optical profiler. Materials and Methods: Dermatologic specimens of 7 cadavers were prepared. Specimens were molded with polyvinyl siloxane and casts prepared with resin, which were scanned using a 3D white light optical profiler. Each scan produced 48 variables, categorized into 3 groups for each location: control, lesion center, and lesion edge. Images of the histopathology slides for suspected seborrheic keratoses were reviewed by a dermatopathologist. Results: The parameters under investigation included border versus center of keratoses, age, sex, lesion location, degree of sun exposure, and cause of death. Although some parameters differ between individuals and age groups, the majority of differences identified between the roughness parameters measured are a result of sex, sun exposure, and histological diagnosis, listed in order of increasing importance. Histological diagnosis provided the most significant, definitive number of individual measurements of areas of roughness in seborrheic keratoses in comparison to other parts of the skin (regardless if those were controls or a different pigmented lesion aside from seborrheic keratoses). Conclusions: This study demonstrates that there are quantifiable patterns of surface textures that can be compared between keratotic and standard normal skin surfaces. These findings suggest this method has the potential to be applied as a noninvasive adjunct to current methods of dermatological diagnosis., Comment: 12 pages, 4 figures

3D integration of GaSb-based gain chips on a silicon photonics platform using micro-transfer printing is demonstrated for the first time. The release process of GaSb coupons, and their transfer for the demonstration of hybrid GaSb/Silicon-photonics on-chip external cavity lasers is reported. A methodology to evaluate the key features of the gain chip coupons, namely the quality of the etched facets and the facet coating deposited using a wafer-level process, is introduced. The characterization provides insight into the fabrication factors limiting the performance of the gain coupons. The level of performance achieved for the transfer printing process offers a solid landmark for the development of photonics integration technology operating at the 2-3 $\mu$m wavelength range. This is instrumental for the deployment of mid-IR photonic integration technology in emerging applications related to gas and biomarker sensing.