Search

This work uses twin 22-year free-running simulations of the Gulf of Mexico hydrodynamics performed with the HYCOM, one considering only ocean dynamics and the other incorporating atmospheric forcing, to study the behavior of the Yucatan Channel transport (YCT), the Loop Current (LC), the Loop Current Eddies (LCEs), their relationships, and the atmospheric forcing effect on them in short (daily) and long (monthly) time scales. A more comprehensive description of the LC intrusion and LCE separations was obtained by considering the upper eastern or western YCT (whose magnitudes are determined by the longitudinal displacements of the Yucatan Current's core), a perspective not evident when considering the upper total YCT; specifically, the eastern YCT provides the most meaningful description of the studied processes. Atmospheric forcing mainly affects the extended stage of the LC by creating a higher dispersion in the YCT and LC circulation values in comparison when considering only ocean dynamics. For the long-term analysis, standardized indexes that integrate the daily values of the eastern YCT and LC circulation in time were used; their temporal propagation and persistence (the changes of their characteristics from short to long time scales) were studied. Intrinsic ocean dynamics produces a persistent YCT and LC intrusion behavior and consistent LCE separation patterns from daily to 5-month scales. The atmospheric forcing effects are more emphasized on the LC intrusion and LCE separations than on the YCT: the YCT persistence is maintained but not that of the LC intrusion. An increased occurrence of LCE separations with low or moderate LC intrusion is expected due to climate change. Using the standardized indexes of the LC metrics to construct a predictive model of the LC intrusion and LCE separations using only current and past LC information is proposed for future research., Comment: 14 pages, 5 figures

Multimodal large language models have fueled progress in image captioning. These models, fine-tuned on vast image datasets, exhibit a deep understanding of semantic concepts. In this work, we show that this ability can be re-purposed for audio captioning, where the joint image-language decoder can be leveraged to describe auditory content associated with image sequences within videos featuring audiovisual content. This can be achieved via multimodal alignment. Yet, this multimodal alignment task is non-trivial due to the inherent disparity between audible and visible elements in real-world videos. Moreover, multimodal representation learning often relies on contrastive learning, facing the challenge of the so-called modality gap which hinders smooth integration between modalities. In this work, we introduce a novel methodology for bridging the audiovisual modality gap by matching the distributions of tokens produced by an audio backbone and those of an image captioner. Our approach aligns the audio token distribution with that of the image tokens, enabling the model to perform zero-shot audio captioning in an unsupervised fashion while keeping the initial image captioning component unaltered. This alignment allows for the use of either audio or audiovisual input by combining or substituting the image encoder with the aligned audio encoder. Our method achieves significantly improved performances in zero-shot audio captioning, compared to existing approaches.

Audio-text models trained via contrastive learning offer a practical approach to perform audio classification through natural language prompts, such as "this is a sound of" followed by category names. In this work, we explore alternative prompt templates for zero-shot audio classification, demonstrating the existence of higher-performing options. First, we find that the formatting of the prompts significantly affects performance so that simply prompting the models with properly formatted class labels performs competitively with optimized prompt templates and even prompt ensembling. Moreover, we look into complementing class labels by audio-centric descriptions. By leveraging large language models, we generate textual descriptions that prioritize acoustic features of sound events to disambiguate between classes, without extensive prompt engineering. We show that prompting with class descriptions leads to state-of-the-art results in zero-shot audio classification across major ambient sound datasets. Remarkably, this method requires no additional training and remains fully zero-shot., Comment: DCASE 2024 - 9th Workshop on Detection and Classification of Acoustic Scenes and Events, Oct 2024, Tokyo, Japan

We present the Deciphering the Interplay between the Interstellar medium, Stars, and the Circumgalactic medium (DIISC) Survey. This survey is designed to investigate the correlations in properties between the circumgalactic medium (CGM), the interstellar medium (ISM), stellar distributions, and young star-forming regions. The galaxies were chosen to have a QSO sightline within 3.5 times the HI radii probing the disk-CGM interface. The sample contains 34 low-redshift galaxies with a median stellar mass of 10$^{10.45}~\rm M_{\odot}$ probed at a median impact parameter of $\rho=55~kpc$. The survey combines ultraviolet spectroscopic data from the Cosmic Origins Spectrograph aboard the Hubble Space Telescope with HI 21 cm hyperfine transition imaging with the Very Large Array (VLA), ultraviolet imaging from Galaxy Evolution Explorer (GALEX), and optical imaging and spectroscopy with the MMT and Vatican Advanced Technology Telescope. We describe the specific goals of the survey, data reduction, high-level data products, and some early results. We present the discovery of a strong inverse correlation, at a confidence level of 99.99%, between Lyman $\alpha$ equivalent width, $\rm W_{Ly\alpha}$, and impact parameter normalized by the HI radius ($\rho/R_{HI}$). We find $\rho/R_{HI}$ to be a better empirical predictor of Lyman $\alpha$ equivalent width than virial radius normalized impact parameter ($\rho/R_{vir}$) or parameterizations combining $\rho,~R_{vir}$, stellar mass, and star formation rate. We conclude that the strong anticorrelation between the Lyman $\alpha$ equivalent width and $\rho/R_{HI}$ indicates that the neutral gas distribution of the CGM is more closely connected to the galaxy's gas disk rather than its stellar and dark matter content., Comment: 29 pages, 12 figures, 4 tables Submitted to AAS Journals

Machine listening systems often rely on fixed taxonomies to organize and label audio data, key for training and evaluating deep neural networks (DNNs) and other supervised algorithms. However, such taxonomies face significant constraints: they are composed of application-dependent predefined categories, which hinders the integration of new or varied sounds, and exhibits limited cross-dataset compatibility due to inconsistent labeling standards. To overcome these limitations, we introduce SALT: Standardized Audio event Label Taxonomy. Building upon the hierarchical structure of AudioSet's ontology, our taxonomy extends and standardizes labels across 24 publicly available environmental sound datasets, allowing the mapping of class labels from diverse datasets to a unified system. Our proposal comes with a new Python package designed for navigating and utilizing this taxonomy, easing cross-dataset label searching and hierarchical exploration. Notably, our package allows effortless data aggregation from diverse sources, hence easy experimentation with combined datasets.

As a part of the Deciphering the Interplay between the Interstellar medium, Stars, and the Circumgalactic medium (DIISC) survey, we investigate indirect evidence of gas inflow into the disk of the galaxy NGC 99. We combine optical spectra from the Binospec spectrograph on the MMT telescope with optical imaging data from the Vatican Advanced Technology Telescope, radio HI 21 cm emission images from the NSF Karl G. Jansky's Very Large Array, and UV spectroscopy from the Cosmic Origins Spectrograph on the Hubble Space Telescope. We measure emission lines (H$\alpha$, H$\beta$, [O III]$\lambda5007$, [N II]$\lambda6583$, and [S II]$\lambda6717,31$) in 26 H II regions scattered about the galaxy and estimate a radial metallicity gradient of $-0.017$ dex kpc$^{-1}$ using the N2 metallicity indicator. Two regions in the sample exhibit an anomalously low metallicity (ALM) of 12+log(O/H) = 8.36 dex, which is $\sim$0.16 dex lower than other regions at that galactocentric radius. They also show a high difference between their HI and H$\alpha$ line of sight velocities on the order of 35 km s$^{-1}$. Chemical evolution modeling indicates gas accretion as the cause of the ALM regions. We find evidence for corotation between the interstellar medium of NGC 99 and Ly$\alpha$ clouds in its circumgalactic medium, which suggests a possible pathway for low metallicity gas accretion. We also calculate the resolved Fundamental Metallicity Relation (rFMR) on sub-kpc scales using localized gas-phase metallicity, stellar mass surface density, and star-formation rate surface density. The rFMR shows a similar trend as that found by previous localized and global FMR relations., Comment: 26 pages, 13 figures, accepted for publication in ApJ

Active components incorporated in materials generate motion by inducing conformational changes in response to external fields. Magnetic fields are particularly interesting as they can actuate materials remotely. Millimeter-sized ferrofluid droplets placed on a solid surface, surrounded by an ambient gas phase, are shown here to migrate under a rotating magnetic field due to the periodic deformation of the liquid-gas interface. This interface wobbling leads to droplet migration with speeds that increase as the amplitude and frequency of the magnetic field increase. In addition to migrating in a controlled manner, we demonstrate the ability of magnetic droplets to clean surface impurities and transport cargo.

Peptide surfactants (PEPS) are studied to capture and retain rare earth elements (REEs) at air-water interfaces to enable REE separations. Peptide sequences, designed to selectively bind REEs, depend crucially on the position of ligands within their binding loop domain. These ligands form a coordination sphere that wraps and retains the cation. We study variants of lanthanide binding tags (LBTs) designed to complex strongly with Tb$^{3+}$. The peptide LBT$^{5-}$ (with net charge -5) is known to bind Tb$^{3+}$ and adsorb with more REE cations than peptide molecules, suggesting that undesired non-specific Coulombic interactions occur. Rheological characterization of interfaces of LBT$^{5-}$ and Tb$^{3+}$ solutions reveal the formation of an interfacial gel. To probe whether this gelation reflects chelation among intact adsorbed LBT$^{5-}$:Tb$^{3+}$ complexes or destruction of the binding loop, we study a variant, LBT$^{3-}$, designed to form net neutral LBT$^{3-}$:Tb$^{3+}$ complexes. Solutions of LBT$^{3-}$ and Tb$^{3+}$ form purely viscous layers in the presence of excess Tb$^{3+}$, indicating that each peptide binds a single REE in an intact coordination sphere. We introduce the variant RR-LBT$^{3-}$ with net charge -3 and anionic ligands outside of the coordination sphere. We find that such exposed ligands promote interfacial gelation. Thus, a nuanced requirement for interfacial selectivity of PEPS is proposed: that anionic ligands outside of the coordination sphere must be avoided to prevent the non-selective recruitment of REE cations. This view is supported by simulation, including interfacial molecular dynamics simulations, and interfacial metadynamics simulations of the free energy landscape of the binding loop conformational space.

To address the global issue of online hate, hate speech detection (HSD) systems are typically developed on datasets from the United States, thereby failing to generalize to English dialects from the Majority World. Furthermore, HSD models are often evaluated on non-representative samples, raising concerns about overestimating model performance in real-world settings. In this work, we introduce NaijaHate, the first dataset annotated for HSD which contains a representative sample of Nigerian tweets. We demonstrate that HSD evaluated on biased datasets traditionally used in the literature consistently overestimates real-world performance by at least two-fold. We then propose NaijaXLM-T, a pretrained model tailored to the Nigerian Twitter context, and establish the key role played by domain-adaptive pretraining and finetuning in maximizing HSD performance. Finally, owing to the modest performance of HSD systems in real-world conditions, we find that content moderators would need to review about ten thousand Nigerian tweets flagged as hateful daily to moderate 60% of all hateful content, highlighting the challenges of moderating hate speech at scale as social media usage continues to grow globally. Taken together, these results pave the way towards robust HSD systems and a better protection of social media users from hateful content in low-resource settings., Comment: ACL 2024 main conference. Data and models available at https://github.com/worldbank/NaijaHate

Traveling wave charges lying on the insulating walls of an electrolyte-filled capillary, give rise to oscillatory modes which vanish when averaged over the period of oscillation. They also give rise to a \emph{zero mode} (a unidirectional, time-independent velocity component) which does not vanish. The latter is a nonlinear effect caused by continuous symmetry-breaking due to the quadratic nonlinearity associated with the electric body force in the time-dependent Stokes equations. In this paper we provide a unified view of the effects arising in traveling wave electroosmosis and establish the universal behavior exhibited by the observables. We show that the incipient velocity profiles are self-similar implying that those obtained with a single experimental configuration, can be employed again to attain further insights without the need of repeating the experiment. Certain results from the literature are recovered as special cases of our formulation and we resolve certain paradoxes having appeared in the past. We present simple theoretical expressions, depending on a single fit parameter, that reproduce these profiles, which could thus provide a rapid test of consistency between our theory and future experiment. The effect becomes more pronounced when reducing the transverse dimension of the system, relative to the velocity direction, and increasing the excitation wavelength, and can therefore be employed for unidirectional transport of electrolytes in thin and long capillaries. General relations, expressing the zero mode velocity in terms of the electric potential and the geometry of the system only, can thus be easily adopted to suit alternative experimental settings.

This study reviews the impact of personalization on human-robot interaction. Firstly, the various strategies used to achieve personalization are briefly described. Secondly, the effects of personalization known to date are discussed. They are presented along with the personalized parameters, personalized features, used technology, and use case they relate to. It is observed that various positive effects have been discussed in the literature while possible negative effects seem to require further investigation.

We study the evolution of opinions on a directed network with community structure. Individuals update their opinions synchronously based on a weighted average of their neighbors' opinions, their own previous opinions, and external media signals. Our model is akin to the popular Friedkin-Johnsen model, and is able to incorporate factors such as stubbornness, confirmation bias, selective exposure, and multiple topics, which are believed to play an important role in the formation of opinions. Our main result shows that, in the large graph limit, the opinion process concentrates around its mean-field approximation for any level of edge density, provided the average degree grows to infinity. Moreover, we show that the opinion process exhibits propagation of chaos. We also give results for the trajectories of individual vertices and the stationary version of the opinion process, and prove that the limits in time and in the size of the network commute. The mean-field approximation is explicit and can be used to quantify consensus and polarization., Comment: Added references

The RHIC interaction rate at sPHENIX will reach around 3 MHz in pp collisions and requires the detector readout to reject events by a factor of over 200 to fit the DAQ bandwidth of 15 kHz. Some critical measurements, such as heavy flavor production in pp collisions, often require the analysis of particles produced at low momentum. This prohibits adopting the traditional approach, where data rates are reduced through triggering on rare high momentum probes. We explore a new approach based on real-time AI technology, adopt an FPGA-based implementation using a custom designed FELIX-712 board with the Xilinx Kintex Ultrascale FPGA, and deploy the system in the detector readout electronics loop for real-time trigger decision., Comment: 7 pages, 5 figures, proceedings for TWEPP 2023 conference, v2: corrected Table 1 numbers

Current state-of-the-art audio analysis systems rely on pre-trained embedding models, often used off-the-shelf as (frozen) feature extractors. Choosing the best one for a set of tasks is the subject of many recent publications. However, one aspect often overlooked in these works is the influence of the duration of audio input considered to extract an embedding, which we refer to as Temporal Support (TS). In this work, we study the influence of the TS for well-established or emerging pre-trained embeddings, chosen to represent different types of architectures and learning paradigms. We conduct this evaluation using both musical instrument and environmental sound datasets, namely OpenMIC, TAU Urban Acoustic Scenes 2020 Mobile, and ESC-50. We especially highlight that Audio Spectrogram Transformer-based systems (PaSST and BEATs) remain effective with smaller TS, which therefore allows for a drastic reduction in memory and computational cost. Moreover, we show that by choosing the optimal TS we reach competitive results across all tasks. In particular, we improve the state-of-the-art results on OpenMIC, using BEATs and PaSST without any fine-tuning., Comment: Copyright 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works

The interaction between passive and active/driven particles has introduced a new way to control colloidal suspension properties from particle aggregation to crystallization. Here, we focus on the hydrodynamic interaction between a single rotational driven particle and a suspension of passive particles near the floor. Using experiments and Stokesian dynamics simulations that account for near-field lubrication, we demonstrate that the flow induced by the driven particle can induce long-ranged rearrangement in a passive suspension. We observe an accumulation of passive particles in front of the driven particle and a depletion of passive particles behind the driven particle. This restructuring generates a pattern that can span a range more than $10$ times of the driven particle's radius. We further show that size scale of the pattern is only a function of the particle's height above the floor., Comment: S.-Y.C. and H.L-R. contributed equally

Dielectric interfaces are crucial to the behavior of charged membranes, from graphene to synthetic and biological lipid bilayers. Understanding electrolyte behavior near these interfaces remains a challenge, especially in the case of rough dielectric surfaces. A lack of analytical solutions consigns this problem to numerical treatments. We report an analytic method for determining electrostatic potentials near curved dielectric membranes in a two-dimensional periodic 'slab' geometry using a periodic summation of Green's functions. This method is amenable to simulating arbitrary groups of charges near surfaces with two-dimensional deformations. We concentrate on one-dimensional undulations. We show that increasing membrane undulation increases the asymmetry of interfacial charge distributions due to preferential ionic repulsion from troughs. In the limit of thick membranes we recover results mimicking those for electrolytes near a single interface. Our work demonstrates that rough surfaces generate charge patterns in electrolytes of charged molecules or mixed-valence ions., Comment: Main article: 25 pages, 6 figures; Supplementary Material: 5 pages, 5 figures

Acid-base equilibria directly influence the functionality and behavior of particles in a system. Due to the ionizing effects of acid-base functional groups, particles will undergo charge exchange. The degree of ionization and their intermolecular and electrostatic interactions are controlled by varying the pH and salt concentration of the solution in a system. Although the pH can be tuned in experiments, it is hard to model this effect using simulations or theoretical approaches. This is due to the difficulty in treating charge regulation and capturing the cooperative effects in a colloidal suspension with Coulombic interaction. In this work, we analyze a suspension of ionizable colloidal particles via Brownian simulations and derive a phase diagram of the system as a function of pH. It is observed that as pH increases, particles functionalized with acid groups change their arrangement from face-centered cubic (FCC) packing to a disordered state. We attribute these transitions to an increase in the degree of charge polydispersity arising from an increase in pH. Our work shows that charge regulation leads to amorphous solids in colloids when the mean nanoparticle charge is sufficiently high., Comment: 15 pages, 12 Figures