Search for: All records

Existing building recognition methods, exemplified by BRAILS, utilize supervised learning to extract information from satellite and street-view images for classification and segmentation. However, each task module requires human-annotated data, hindering the scalability and robustness to regional variations and annotation imbalances. In response, we propose a new zero-shot workflow for building attribute extraction that utilizes large-scale vision and language models to mitigate reliance on external annotations. The proposed workflow contains two key components: image-level captioning and segment-level captioning for the building images based on the vocabularies pertinent to structural and civil engineering. These two components generate descriptive captions by computing feature representations of the image and the vocabularies, and facilitating a semantic match between the visual and textual representations. Consequently, our framework offers a promising avenue to enhance AI-driven captioning for building attribute extraction in the structural and civil engineering domains, ultimately reducing reliance on human annotations while bolstering performance and adaptability. 

We first introduce the notion of meta-rank for a 2-parameter persistence module, an invariant that captures the information behind images of morphisms between 1D slices of the module. We then define the meta-diagram of a 2-parameter persistence module to be the Möbius inversion of the meta-rank, resulting in a function that takes values from signed 1-parameter persistence modules. We show that the meta-rank and meta-diagram contain information equivalent to the rank invariant and the signed barcode. This equivalence leads to computational benefits, as we introduce an algorithm for computing the meta-rank and meta-diagram of a 2-parameter module M indexed by a bifiltration of n simplices in O(n^3) time. This implies an improvement upon the existing algorithm for computing the signed barcode, which has O(n^4) time complexity. This also allows us to improve the existing upper bound on the number of rectangles in the rank decomposition of M from O(n^4) to O(n^3). In addition, we define notions of erosion distance between meta-ranks and between meta-diagrams, and show that under these distances, meta-ranks and meta-diagrams are stable with respect to the interleaving distance. Lastly, the meta-diagram can be visualized in an intuitive fashion as a persistence diagram of diagrams, which generalizes the well-understood persistent diagram in the 1-parameter setting. 

Although many substorm‐related observations have been made, we still have limited insight into propagation of the plasma and field perturbations in Pi2 frequencies (∼7–25 mHz) in association with substorm aurora, particularly from the auroral source region in the inner magnetosphere to the ground. In this study, we present conjugate observations of a substorm brightening aurora using an all‐sky camera and an inner‐magnetospheric satellite Arase atL ∼ 5. A camera at Gakona (62.39°N, 214.78°E), Alaska, observed a substorm auroral brightening on 28 December 2018, and the footprint of the satellite was located just equatorward of the aurora. Around the timing of the auroral brightening, the satellite observed a series of quasi‐periodic variations in the electric and magnetic fields and in the energy flux of electrons and ions. We demonstrate that the diamagnetic variations of thermal pressure and medium‐energy ion energy flux in the inner magnetosphere show approximately one‐to‐one correspondence with the oscillations in luminosity of the substorm brightening aurora and high‐latitudinal Pi2 pulsations on the ground. We also found their anti‐correlation with low‐energy electrons. Cavity‐type Pi2 pulsations were observed at mid‐ and low‐latitudinal stations. Based on these observations, we suggest that a wave phenomenon in the substorm auroral source region, like ballooning type instability, play an important role in the development of substorm and related auroral brightening and high‐latitude Pi2, and that the variation of the auroral luminosity was directly driven by keV electrons which were modulated by Alfven waves in the inner magnetosphere.

 

Abstract Loss functions with a large number of saddle points are one of the major obstacles for training modern machine learning (ML) models efficiently. First-order methods such as gradient descent (GD) are usually the methods of choice for training ML models. However, these methods converge to saddle points for certain choices of initial guesses. In this paper, we propose a modification of the recently proposed Laplacian smoothing gradient descent (LSGD) [Osher et al., arXiv:1806.06317 ], called modified LSGD (mLSGD), and demonstrate its potential to avoid saddle points without sacrificing the convergence rate. Our analysis is based on the attraction region, formed by all starting points for which the considered numerical scheme converges to a saddle point. We investigate the attraction region’s dimension both analytically and numerically. For a canonical class of quadratic functions, we show that the dimension of the attraction region for mLSGD is $\lfloor (n-1)/2\rfloor$ , and hence it is significantly smaller than that of GD whose dimension is $n-1$ . 

Abstract The Voyager 2 crossing of the termination shock indicated that most of the upstream energy from the thermal solar wind ions was transferred to pickup ions (PUIs) and other energetic particles downstream of the shock. We use hybrid simulations at the termination shock for the Voyager 2, flank, and tail directions to evaluate the distributions of different ion species downstream of the shock over the energy range of 0.52–55 keV. Here, we extend the work of Gkioulidou et al., which showed an energy-dependent discrepancy between modeled and energetic neutral atom (ENA) observations, and fit distributions to a hybrid model to show that a population of PUIs accelerated via diffusive shock acceleration (DSA) to become low-energy anomalous cosmic rays (ACRs) can bridge the gap between modeled and observed ENA fluxes. Our results with the inclusion of DSA via hybrid fitting give entirely new and novel evidence that DSA at the termination shock is likely to be an important physical process. These ACRs carry a significant fraction of the energy density at the termination shock (22%, 13%, and 19% in the Voyager 2, flank, and tail directions, respectively). Using these ACRs in global ENA modeling of the heliosphere from 0.52 to 55 keV, we find that scaling factors as large as 1.8–2.5 are no longer required to match ENA observations at energies of ∼1–4 keV. Large discrepancies between modeled and observed ENAs only remain over energies of 4–20 keV, indicating that there may be a further acceleration mechanism in the heliosheath at these energies. 

In this paper, we propose MetaMobi, a novel spatio-temporal multi-dots connectivity-aware modeling and Meta model update approach for crowd Mobility learning. MetaMobi analyzes real-world Wi-Fi association data collected from our campus wireless infrastructure, with the goal towards enabling a smart connected campus. Specifically, MetaMobi aims at addressing the following two major challenges with existing crowd mobility sensing system designs: (a) how to handle the spatially, temporally, and contextually varying features in large-scale human crowd mobility distributions; and (b) how to adapt to the impacts of such crowd mobility patterns as well as the dynamic changes in crowd sensing infrastructures. To handle the first challenge, we design a novel multi-dots connectivity-aware learning approach, which jointly learns the crowd flow time series of multiple buildings with fusion of spatial graph connectivities and temporal attention mechanisms. Furthermore, to overcome the adaptivity issues due to changes in the crowd sensing infrastructures (e.g., installation of new ac- cess points), we further design a novel meta model update approach with Bernoulli dropout, which mitigates the over- fitting behaviors of the model given few-shot distributions of new crowd mobility datasets. Extensive experimental evaluations based on the real-world campus wireless dataset (including over 76 million Wi-Fi association and disassociation records) demonstrate the accuracy, effectiveness, and adaptivity of MetaMobi in forecasting the campus crowd flows, with 30% higher accuracy compared to the state-of-the-art approaches. 

The 1D power spectrum P1D of the Ly α forest provides important information about cosmological and astrophysical parameters, including constraints on warm dark matter models, the sum of the masses of the three neutrino species, and the thermal state of the intergalactic medium. We present the first measurement of P1D with the quadratic maximum likelihood estimator (QMLE) from the Dark Energy Spectroscopic Instrument (DESI) survey early data sample. This early sample of 54 600 quasars is already comparable in size to the largest previous studies, and we conduct a thorough investigation of numerous instrumental and analysis systematic errors to evaluate their impact on DESI data with QMLE. We demonstrate the excellent performance of the spectroscopic pipeline noise estimation and the impressive accuracy of the spectrograph resolution matrix with 2D image simulations of raw DESI images that we processed with the DESI spectroscopic pipeline. We also study metal line contamination and noise calibration systematics with quasar spectra on the red side of the Ly α emission line. In a companion paper, we present a similar analysis based on the Fast Fourier Transform estimate of the power spectrum. We conclude with a comparison of these two approaches and discuss the key sources of systematic error that we need to address with the upcoming DESI Year 1 analysis.