arXiv subjects

Siyang Li

Publications and source records attributed to Siyang Li.

At least 55 records · Page 3Linked to original sources

arXiv2024

The Tip of the Red Giant Branch Distance Ladder and the Hubble Constant

While the tip of the red giant branch (TRGB) has been used as a distance indicator since the early 1990's, its application to measure the Hubble Constant as a primary distance indicator occurred only recently. The TRGB is also currently at an interesting crossroads as results from the James Webb Space Telescope (JWST) are beginning to emerge. In this chapter, we provide a review of the TRGB as it is used to measure the Hubble constant. First, we provide an essential review of the physical and observational basis of the TRGB as well as providing a summary for its use for measuring the Hubble Constant. More attention is then given is then given to recent, but still pre-JWST, developments, including new calibrations and developments with algorithms. We also address challenges that arise while measuring a TRGB-based Hubble Constant. We close by looking forward to the exciting prospects from telescopes such as JWST and Gaia.

astro-ph.CO↗

arXiv2024

Reconnaissance with JWST of the J-region Asymptotic Giant Branch in Distance Ladder Galaxies: From Irregular Luminosity Functions to Approximation of the Hubble Constant

We study stars in the J-regions of the asymptotic giant branch (JAGB) of near-infrared color magnitude diagrams in the maser host NGC 4258 and 4 hosts of 6 Type Ia supernovae (SN Ia): NGC 1448, NGC 1559, NGC 5584, and NGC 5643. These clumps of stars are readily apparent near $1.0<F150W-F277W<1.5$ and $m_{F150W}$=22-25 mag with James Webb Space Telescope NIRCam photometry. Various methods have been proposed to assign an apparent reference magnitude for this recently proposed standard candle, including the mode, median, sigma-clipped mean or a modeled luminosity function parameter. We test the consistency of these by measuring intra-host variations, finding differences of up to $\sim$0.2 mag that significantly exceed statistical uncertainties. Brightness differences appear intrinsic, and are further amplified by the non-uniform shape of the JAGB luminosity function, also apparent in the LMC and SMC. We follow a 'many methods' approach to consistently measure JAGB magnitudes and distances to the SN Ia host sample calibrated by NGC 4258. We find broad agreement with distances measured from Cepheids, tip of the red giant branch (TRGB), and Miras. However, the SN host mean distance estimated via the JAGB method necessary to estimate $H_0$ differs by $\sim$0.19 mag amongst the above definitions, a result of different levels of luminosity function asymmetry. The methods yield a full range of $71-78$ km s$^{-1}$ Mpc$^{-1}$, i.e., a fiducial result of $H_0=74.7\pm2.1$ (stat) $\pm$2.3 (sys) ($\pm$3.1 if combined in quadrature) km s$^{-1}$ Mpc$^{-1}$, with systematic errors limited by the differences in methods. Future work may seek to further standardize and refine this promising tool, making it more competitive with established distance indicators.

astro-ph.CO↗

arXiv2024

DiffPLF: A Conditional Diffusion Model for Probabilistic Forecasting of EV Charging Load

Due to the vast electric vehicle (EV) penetration to distribution grid, charging load forecasting is essential to promote charging station operation and demand-side management.However, the stochastic charging behaviors and associated exogenous factors render future charging load patterns quite volatile and hard to predict. Accordingly, we devise a novel Diffusion model termed DiffPLF for Probabilistic Load Forecasting of EV charging, which can explicitly approximate the predictive load distribution conditioned on historical data and related covariates. Specifically, we leverage a denoising diffusion model, which can progressively convert the Gaussian prior to real time-series data by learning a reversal of the diffusion process. Besides, we couple such diffusion model with a cross-attention-based conditioning mechanism to execute conditional generation for possible charging demand profiles. We also propose a task-informed fine-tuning technique to better adapt DiffPLF to the probabilistic time-series forecasting task and acquire more accurate and reliable predicted intervals. Finally, we conduct multiple experiments to validate the superiority of DiffPLF to predict complex temporal patterns of erratic charging load and carry out controllable generation based on certain covariate. Results demonstrate that we can attain a notable rise of 39.58% and 49.87% on MAE and CRPS respectively compared to the conventional method.

cs.LG↗

arXiv2024

DiffCharge: Generating EV Charging Scenarios via a Denoising Diffusion Model

Recent proliferation of electric vehicle (EV) charging events has brought prominent stress over power grid operation. Due to the stochastic and volatile EV charging behaviors, the induced charging loads are extremely uncertain, posing modeling and control challenges for grid operators and charging management. Generating EV charging scenarios would aid via synthesizing a myriad of realistic charging scenarios. To this end, we propose a novel denoising Diffusion-based Charging scenario generation model DiffCharge, which is capable of generating a broad variety of realistic EV charging profiles with distinctive temporal properties. It is able to progressively convert the simply known Gaussian noise to genuine charging time-series data, by learning a parameterized reversal of a forward diffusion process. Besides, we leverage the multi-head self-attention and prior conditions to capture the temporal correlations and unique information associated with EV or charging station types in real charging profiles. Moreover, We demonstrate the superiority of DiffCharge on extensive real-world charging datasets, as well as the efficacy on EV integration in power distribution grids.

eess.SY↗

arXiv2023

Large Foundation Models for Power Systems

Foundation models, such as Large Language Models (LLMs), can respond to a wide range of format-free queries without any task-specific data collection or model training, creating various research and application opportunities for the modeling and operation of large-scale power systems. In this paper, we outline how such large foundation model such as GPT-4 are developed, and discuss how they can be leveraged in challenging power and energy system tasks. We first investigate the potential of existing foundation models by validating their performance on four representative tasks across power system domains, including the optimal power flow (OPF), electric vehicle (EV) scheduling, knowledge retrieval for power engineering technical reports, and situation awareness. Our results indicate strong capabilities of such foundation models on boosting the efficiency and reliability of power system operational pipelines. We also provide suggestions and projections on future deployment of foundation models in power system applications.

eess.SY↗

arXiv2023

Tunable topological phase transition in soft Rayleigh beam system with imperfect interfaces

Acoustic metamaterials, particularly the topological insulators, exhibit exceptional wave characteristics that have sparked considerable research interest. The study of imperfect interfaces affect is of significant importance for the modeling of wave propagation behavior in topological insulators. This paper models a soft Rayleigh beam system with imperfect interfaces, and investigates its topological phase transition process tuned by mechanical loadings. The model reveals that the topological phase transition process can be observed by modifying the distance between imperfect interfaces in the system. When a uniaxial stretch is applied, the topological phase transition points for longitudinal waves decrease within a limited frequency range, while they increase within a larger frequency scope for transverse waves. Enhancing the rigidity of the imperfect interfaces also enables shifting of the topological phase transition point within a broader frequency range for longitudinal waves and a confined range for transverse waves. The transition of topologically protected interface modes in the transmission performance of a twenty-cell system is verified, which include altering frequencies, switching from interface mode to edge mode. Overall, this study provides a new approach and guideline for controlling topological phase transition in composite and soft phononic crystal systems.

cond-mat.soft↗

arXiv2023

Standardized Luminosity of the Tip of the Red Giant Branch utilizing Multiple Fields in NGC 4258 and the CATs Algorithm

The Tip of the Red Giant Branch provides a luminous standard candle for calibrating distance ladders that reach Type Ia supernova (SN Ia) hosts. However, recent work reveals that tip measurements vary at the $\sim$ 0.1 mag level for different stellar populations and locations within a host, which may lead to inconsistencies along the distance ladder. We pursue a calibration of the tip using 11 Hubble Space Telescope fields around the maser host, NGC 4258, that is consistent with SN Ia hosts by standardizing tip measurements via their contrast ratios. We find $F814W$-band tips that exhibit a full 0.3 mag range and 0.1 mag dispersion. We do not find any correlation between HI column density and the apparent tip to 0.04 $\pm$ 0.03 mag/cm$^{-2}$. We search for a tip-contrast relation (TCR) and measure the TCR within the fields of NGC 4258 of $-0.015\pm0.008$ mag/$R$, where $R$ is the contrast ratio. This value is consistent with the TCR originally discovered in the GHOSTS sample (Wu et al. 2022) of $-0.023\pm0.005$ mag/R. Combining these measurements, we find a global TCR of $-0.021\pm0.004$ mag/R and a calibration of $M_I^{TRGB} = -4.025 \pm 0.035 - (R-4)\times0.021$ mag. We also use stellar models to simulate single age and metallicity stellar populations with [Fe/H] from $-2.0$ to $-0.7$ and ages from 3 Gyr to 12 Gyr and reconstruct the global TCR found here to a factor of $\sim$ 2. This work is combined in a companion analysis with tip measurements of nearby SN Ia hosts to measure $H_0$.

astro-ph.GA↗

arXiv2023

A Gaia Data Release 3 View on the Tip of the Red Giant Branch Luminosity

The tip of the red giant branch (TRGB) is a standard candle that can be used to help refine the determination of the Hubble constant. $Gaia$ Data Release 3 (DR3) provides synthetic photometry constructed from low-resolution BP/RP spectra for Milky Way field stars that can be used to directly calibrate the luminosity of the TRGB in the Johnson-Cousins I band, where the TRGB is least sensitive to metallicity. We calibrate the TRGB luminosity using a two-dimensional maximum likelihood algorithm with field stars and $Gaia$ synthetic photometry and parallaxes. For a high-contrast and low-contrast break (characterized by the values of the contrast parameter $ R$ or the magnitude of the break $ β$), we find $M^{TRGB}_I$ =$-4.02$ and $-3.92$ mag respectively, or a midpoint of $-3.970$ $^{+0.042} _{-0.024}$ (sys) $\pm$ $0.062$ (stat) mag. This measurement improves upon the TRGB measurement from Li et al. (2022), as the higher precision photometry based on $ Gaia $ DR3 allows us to constrain two additional free parameters of the luminosity function. We also investigate the possibility of using $Gaia$ DR3 synthetic photometry to calibrate the TRGB luminosity with $ω$ Centauri, but find evidence of blending within the inner region for cluster member photometry that precludes accurate calibration with $Gaia$ DR3 photometry. We instead provide an updated TRGB measurement of $m^{TRGB}_I$ = $ 9.82 \pm 0.04$ mag in $ω$ Centauri using ground-based photometry from the most recent version of the database described in Stetson et al. (2019), which gives $M^{TRGB}_I$ = $-3.97$ $\pm$ $0.04$ (stat) $\pm$ 0.10 (sys) mag when tied to the $Gaia$ EDR3 parallax distance from the consensus of Vasiliev & Baumgardt (2021), Soltis et al. (2021), and Maíz Apellániz et al. (2022a).

astro-ph.GA↗

arXiv2023

A 1.3% distance to M33 from HST Cepheid photometry

We present a low-dispersion period-luminosity relation (PL) based on 154 Cepheids in Messier 33 (M33) with Hubble Space Telescope (HST) photometry from the PHATTER survey. Using high-quality ground-based light curves, we recover Cepheid phases and amplitudes for multi-epoch HST data and we perform template fitting to derive intensity-averaged mean magnitudes. HST observations in the SH0ES near-infrared Wesenheit system significantly reduce the effect of crowding relative to ground-based data, as seen in the final PL scatter of $σ$ = 0.11 mag. We adopt the absolute calibration of the PL based on HST observations in the Large Magellanic Cloud (LMC) and a distance derived using late-type detached eclipsing binaries to obtain a distance modulus for M33 of $μ$ = 24.622 $\pm$ 0.030 mag (d = 840 $\pm$ 11 kpc), a best-to-date precision of 1.3%. We find very good agreement with past Cepheid-based measurements. Several TRGB estimates bracket our result while disagreeing with each other. Finally, we show that the flux contribution from star clusters hosting Cepheids in M33 does not impact the distance measurement and we find only 3.7% of the sample is located in (or nearby) young clusters. M33 offers one of the best sites for the cross-calibration of many primary distance indicators. Thus, a precise independent geometric determination of its distance would provide a valuable new anchor to measure the Hubble constant.

astro-ph.CO↗

arXiv2022

A Maximum Likelihood Calibration of the Tip of the Red Giant Branch Luminosity from High Latitude Field Giants using Gaia Early Data Release 3 Parallaxes

The calibration of the tip of the red giant branch (TRGB) in the $I$-band has a direct role in determinations of the Hubble constant, a subject of recent interest due to the discrepancy between direct and indirect estimates of its value. We present a maximum likelihood (ML) method designed to obtain an independent calibration of the brightness of TRGB using $Gaia$ parallaxes from the Early Data Release 3 (EDR3) of Milky Way field Giants at high Galactic latitude. We adopt simple parameterizations for the Milky Way stellar luminosity function and density law and and optimize the likelihood of the observed sample as a function of those parameters. Using parameters to partially constrain the luminosity function from other galaxies similar to the Milky Way for which high quality TRGB data are available, we find values of the TRGB magnitude of $ M_I^{TRGB} = -3.91 \pm 0.05 $ (stat) $ \pm 0.09 $ (sys) mag, where the systematic uncertainty covers the range of shape parameters found in our Milky Way sample and in reference galaxies. While APASS Data Release 9 all-sky photometry is insufficient to provide a reliable constraint on the shape of the Milky Way luminosity function, we estimate that the photometry from $Gaia$ Data Release 3 (mid-2022) will allow better constraints on the shape, and lower statistical uncertainties on the tip by a factor of 3. With expected releases of improved parallax measurements from $Gaia$, the method of calibrating the TRGB luminosity from field Giants is expected to reach $\sim$ 0.01 mag uncertainty, which is an important step toward a precise TRGB-based determination of the Hubble constant.

astro-ph.GA↗

arXiv2022

Facial Affect Analysis: Learning from Synthetic Data & Multi-Task Learning Challenges

Facial affect analysis remains a challenging task with its setting transitioned from lab-controlled to in-the-wild situations. In this paper, we present novel frameworks to handle the two challenges in the 4th Affective Behavior Analysis In-The-Wild (ABAW) competition: i) Multi-Task-Learning (MTL) Challenge and ii) Learning from Synthetic Data (LSD) Challenge. For MTL challenge, we adopt the SMM-EmotionNet with a better ensemble strategy of feature vectors. For LSD challenge, we propose respective methods to combat the problems of single labels, imbalanced distribution, fine-tuning limitations, and choice of model architectures. Experimental results on the official validation sets from the competition demonstrated that our proposed approaches outperformed baselines by a large margin. The code is available at https://github.com/sylyoung/ABAW4-HUST-ANT.

cs.LG↗

arXiv2021

Low-Resource Machine Translation Training Curriculum Fit for Low-Resource Languages

We conduct an empirical study of neural machine translation (NMT) for truly low-resource languages, and propose a training curriculum fit for cases when both parallel training data and compute resource are lacking, reflecting the reality of most of the world's languages and the researchers working on these languages. Previously, unsupervised NMT, which employs back-translation (BT) and auto-encoding (AE) tasks has been shown barren for low-resource languages. We demonstrate that leveraging comparable data and code-switching as weak supervision, combined with BT and AE objectives, result in remarkable improvements for low-resource languages even when using only modest compute resources. The training curriculum proposed in this work achieves BLEU scores that improve over supervised NMT trained on the same backbone architecture by +12.2 BLEU for English to Gujarati and +3.7 BLEU for English to Kazakh, showcasing the potential of weakly-supervised NMT for the low-resource languages. When trained on supervised data, our training curriculum achieves a new state-of-the-art result on the Somali dataset (BLEU of 29.3 for Somali to English). We also observe that adding more time and GPUs to training can further improve performance, which underscores the importance of reporting compute resource usage in MT research.

cs.CL↗

arXiv2021

A sub-2% Distance to M31 from Photometrically Homogeneous Near-Infrared Cepheid Period-Luminosity Relations Measured with the Hubble Space Telescope

We present Period-Luminosity Relations (PLRs) for 55 Cepheids in M31 with periods ranging from 4 to 78 days observed with the Hubble Space Telescope (HST) using the same three-band photometric system recently used to calibrate their luminosities. Images were taken with the Wide Field Camera 3 in two optical filters (F555W and F814W) and one near-infrared filter (F160W) using the Drift And SHift (DASH) mode of operation to significantly reduce overheads and observe widely-separated Cepheids in a single orbit. We include additional F160W epochs for each Cepheid from the Panchromatic Hubble Andromeda Treasury (PHAT) and use light curves from the Panoramic Survey Telescope and Rapid Response System of the Andromeda galaxy (PAndromeda) project to determine mean magnitudes. Combined with a 1.28$\%$ absolute calibration of Cepheid PLRs in the Large Magellanic Cloud from Riess et al. (2019a) in the same three filters, we find a distance modulus to M31 of $μ_0$ = 24.407 $\pm$ 0.032, corresponding to 761 $\pm$ 11 kpc and 1.49$\%$ uncertainty including all error sources, the most precise determination of its distance to date. We compare our results to past measurements using Cepheids and the Tip of the Red Giant Branch (TRGB). This study also provides the groundwork for turning M31 into a precision anchor galaxy in the cosmic distance ladder to measure the Hubble constant together with efforts to measure a fully geometric distance to M31.

astro-ph.CO↗

arXiv2021

The surprising impact of mask-head architecture on novel class segmentation

Instance segmentation models today are very accurate when trained on large annotated datasets, but collecting mask annotations at scale is prohibitively expensive. We address the partially supervised instance segmentation problem in which one can train on (significantly cheaper) bounding boxes for all categories but use masks only for a subset of categories. In this work, we focus on a popular family of models which apply differentiable cropping to a feature map and predict a mask based on the resulting crop. Under this family, we study Mask R-CNN and discover that instead of its default strategy of training the mask-head with a combination of proposals and groundtruth boxes, training the mask-head with only groundtruth boxes dramatically improves its performance on novel classes. This training strategy also allows us to take advantage of alternative mask-head architectures, which we exploit by replacing the typical mask-head of 2-4 layers with significantly deeper off-the-shelf architectures (e.g. ResNet, Hourglass models). While many of these architectures perform similarly when trained in fully supervised mode, our main finding is that they can generalize to novel classes in dramatically different ways. We call this ability of mask-heads to generalize to unseen classes the strong mask generalization effect and show that without any specialty modules or losses, we can achieve state-of-the-art results in the partially supervised COCO instance segmentation benchmark. Finally, we demonstrate that our effect is general, holding across underlying detection methodologies (including anchor-based, anchor-free or no detector at all) and across different backbone networks. Code and pre-trained models are available at https://git.io/deepmac.

cs.CV↗

arXiv2020

Climate change adaption in Chinese ancient architecture

As an important symbol of civilization and culture, architectures originally were built for sheltering human beings from weather disasters and therefore should be affected by climate change, particularly the associated change in the occurrence of extreme weather events. However, although meteorology has been considered as a factor in modern architecture design, it remains unclear whether and how the ancients adapted to climate change from the perspective of architecture design, particularly on a millennium time scale. Here we show that the periodic change and trend of the roof slope of ancient architectures in northern China clearly demonstrate the climate change adaptation over the past thousand years. We show that the snowfall fluctuation caused by the paleo-climate change was an essential driving factor within the roof modification process of ancient Chinese timber architecture. Our study indicates that climate change may act as a much more important factor on human behaviour than we ever thought.

physics.ao-ph↗

arXiv2020

On-sky silicon photomultiplier detector performance measurements for millisecond to sub-microsecond optical source variability studies

In our Ultra-Fast Astronomy (UFA) program, we aim to improve measurements of variability of astronomical targets on millisecond and shorter time scales. In this work, we present initial on-sky measurements of the performance of silicon photomultiplier detectors (SiPMs) for UFA. We mounted two different SiPMs at the focal plane of the 0.7-meter aperture Nazarbayev University Transient Telescope at the Assy-Turgen Astrophysical Observatory (NUTTelA-TAO), with no filter in front of the detector. The $3mm\times3mm$ SiPM single-channel detectors have a field of view of $2.2716'\times2.2716'$. During the nights of 2019 October 28-29, we measured sky background, bright stars, and an artificial source with a 100Hz flashing frequency. We compared detected SiPM counts with Gaia satellite G-band flux values to show that our SiPMs have a linear response. With our two SiPMs (models S14520-3050VS and S14160-3050HS), we measured a dark current of $\sim$130 and $\sim$85 kilo counts per second (kcps), and a sky background of $\sim$201 and $\sim$203 kcps, respectively. We measured an intrinsic crosstalk of 10.34$\%$ and 10.52$\%$ and derived a 5$σ$ sensitivity of 13.9 and 14.0 Gaia G-band magnitude for 200ms exposures, for the two detectors respectively. For a 10 $μ$s window, and allowing a false alarm rate of once per 100 nights, we derived a sensitivity of 22 detected photons, or 6 Gaia G-band magnitudes. For nanosecond timescales, our detection is limited by crosstalk to 12 detected photons, which corresponds to a fluence of $\sim$155 photons per square meter.

astro-ph.IM↗

arXiv2020

Characterization of a high efficiency silicon photomultiplier for millisecond to sub-microsecond astrophysical transient searches

We characterized the S14160-3050HS Multi-Pixel Photon Counter (MPPC), a high efficiency, single channel silicon photomultiplier manufactured by Hamamatsu Photonics K.K. All measurements were performed at a room temperature of (23.0 $\pm$ 0.3) $^{\circ}$C. We obtained an I-V curve and used relative derivatives to find a breakdown voltage of 38.88 V. At a 3 V over voltage, we find a dark count rate of 1.08 MHz, crosstalk probability of 21 $\%$, photon detection efficiency of 55 $\%$ at 450 nm, and saturation at 1.0x10$^{11}$ photons per second. The S14160-3050HS MPPC is a candidate detector for the Ultra-Fast Astronomy (UFA) telescope which will characterize the optical (320 nm - 650 nm) sky in the millisecond to sub-microsecond timescales using two photon counting arrays operated in coincidence on the 0.7 meter Nazarbayev University Transient Telescope at the Assy-Turgen Astrophysical Observatory (NUTTelA-TAO) located near Almaty, Kazakhstan. We discuss advantages and disadvantages of using the S14160-3050HS MPPC for the UFA telescope and future ground-based telescopes in sub-second time domain astrophysics.

astro-ph.IM↗

arXiv2019

Program objectives and specifications for the Ultra-Fast Astronomy observatory

We present program objectives and specifications for the first generation Ultra-Fast Astronomy (UFA) observatory which will explore a new astrophysical phase space by characterizing the variability of the optical (320 nm - 650 nm) sky in the millisecond to nanosecond timescales. One of the first objectives of the UFA observatory will be to search for optical counterparts to fast radio bursts (FRB) that can be used to identify the origins of FRB and probe the epoch of reionization and baryonic matter in the interstellar and intergalactic mediums. The UFA camera will consist of two single-photon resolution fast-response detector 16x16 arrays operated in coincidence mounted on the 0.7 meter Nazarbayev University Transient Telescope at the Assy-Turgen Astrophysical Observatory (NUTTelA-TAO) located near Almaty, Kazakhstan. We are currently developing two readout systems that can measure down to the microsecond and nanosecond timescales and characterizing two silicon photomultipliers (SiPM) and one photomultiplier tube (PMT) to compare the detectors for the UFA observatory and astrophysical observations in general.

astro-ph.IM↗