Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userFlare Statistics for Young Stars from a Convolutional Neural Network Analysis of $textit{TESS}$ Data
64
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
All-sky photometric time-series missions have allowed for the monitoring of thousands of young ($t_{rm age} < 800$Myr) to understand the evolution of stellar activity. Here we developed a convolutional neural network (CNN), $texttt{stella}$, specifically trained to find flares in $textit{Transiting Exoplanet Survey Satellite}$ ($textit{TESS}$) short-cadence data. We applied the network to 3200 young stars to evaluate flare rates as a function of age and spectral type. The CNN takes a few seconds to identify flares on a single light curve. We also measured rotation periods for 1500 of our targets and find that flares of all amplitudes are present across all spot phases, suggesting high spot coverage across the entire surface. Additionally, flare rates and amplitudes decrease for stars $t_{rm age} > 50$Myr across all temperatures $T_{rm eff} geq 4000$K, while stars from $2300 leq T_{rm eff} < 4000$K show no evolution across 800 Myr. Stars of $T_{rm eff} leq 4000$K also show higher flare rates and amplitudes across all ages. We investigate the effects of high flare rates on photoevaporative atmospheric mass loss for young planets. In the presence of flares, planets lose 4-7% more atmosphere over the first 1 Gyr. $texttt{stella}$ is an open-source Python tool-kit hosted on GitHub and PyPI.
rate research
Read More
Phased flaring, or the periodic occurrence of stellar flares, may probe electromagnetic star-planet interaction (SPI), binary interaction, or magnetic conditions in spots. For the first time, we explore flare periodograms for a large sample of flare stars to identify periodicity due to magnetic interactions with orbiting companions, magnetic reservoirs, or rotational phase. Previous large surveys have explored periodicity at the stellar rotation period, but we do not assume periods must correspond with rotation in this work. Two min TESS light curves of 284 cool stars are searched for periods from 1-10 d using two newly-developed periodograms. Because flares are discrete events in noisy and incomplete data, typical periodograms are not well-suited to detect phased flaring. We construct and test a new Bayesian likelihood periodogram and a modified Lomb-Scargle periodogram. We find 6 candidates with a false-alarm probability below 1%. Three targets are >3-sigma detections of flare periodicity; the others are plausible candidates which cannot be individually confirmed. Periods range from 1.35 to 6.7 d and some, but not all, correlate with the stellar rotation period or its 1/2 alias. Periodicity from 2 targets may persist from TESS Cycle 1 into Cycle 3. The periodicity does not appear to persist for the others. Long-term changes in periodicity may result from the spot evolution observed from each candidate, which suggests magnetic conditions play an important role in sustaining periodicity.
The study is devoted to search for flare stars among confirmed members of Galactic open clusters using high-cadence photometry from {it TESS} mission. We analyzed 957 high-cadence light curves of members from 136 open clusters. As a result, 56 flare stars were found, among them 8 hot B-A type objects. Of all flares, 63% were detected in a sample of cool stars ($T_{rm eff}<5000$~K), and 29% -- in stars of spectral type G, while 23% in K-type stars and approximately 34% of all detected flares are in M-type stars. Using the FLATWRM (FLAre deTection With Ransac Method) flare finding algorithm, we estimated parameters of flares and rotation period of detected flare stars. The flare with the largest amplitude appears on the M3 type EQ,Cha star. Statistical analysis did not reveal any direct correlation between ages, rotation periods and flaring activity.
We have identified a quadruple system with two close eclipsing binaries in TESS data. The object is unresolved in Gaia and appears as a single source at parallax 1.08~$pm$0.01 mas. Both binaries have observable primary and secondary eclipses and were monitored throughout TESS Cycle 1 (sectors 1-13), falling within the TESS Continuous Viewing Zone. In one eclipsing binary (P = 5.488 d), the smaller star is completely occluded by the larger star during the secondary eclipse; in the other (P = 5.674 d) both eclipses are grazing. Using these data, spectroscopy, speckle photometry, SED analysis and evolutionary stellar tracks, we have constrained the masses and radii of the four stars in the two eclipsing binaries. The Li I EW indicates an age of 10-50 Myr and, with an outer period of $858^{+7}_{-5}$ days, our analysis indicates this is one of the most compact young 2+2 quadruple systems known.
Due to the ever-expanding volume of observed spectroscopic data from surveys such as SDSS and LAMOST, it has become important to apply artificial intelligence (AI) techniques for analysing stellar spectra to solve spectral classification and regression problems like the determination of stellar atmospheric parameters Teff, log g, [Fe/H]. We propose an automated approach for the classification of stellar spectra in the optical region using Convolutional Neural Networks. Traditional machine learning (ML) methods with shallow architecture (usually up to 2 hidden layers) have been trained for these purposes in the past. However, deep learning methods with a larger number of hidden layers allow the use of finer details in the spectrum which results in improved accuracy and better generalisation. Studying finer spectral signatures also enables us to determine accurate differential stellar parameters and find rare objects. We examine various machine and deep learning algorithms like Artificial Neural Networks (ANN), Random Forest (RF), and Convolutional Neural Network (CNN) to classify stellar spectra using the Jacoby Atlas, ELODIE and MILES spectral libraries as training samples. We test the performance of the trained networks on the Indo-U.S. Library of Coude Feed Stellar Spectra (CFLIB). We show that using convolutional neural networks, we are able to lower the error up to 1.23 spectral sub-classes as compared to that of 2 sub-classes achieved in the past studies with ML approach. We further apply the trained model to classify stellar spectra retrieved from the SDSS database with SNR>20.
In this paper, we propose a new data augmentation strategy named Thumbnail, which aims to strengthen the networks capture of global features. We get a generated image by reducing an image to a certain size, which is called as the thumbnail, and pasting it in the random position of the original image. The generated image not only retains most of the original image information but also has the global information in the thumbnail. Furthermore, we find that the idea of thumbnail can be perfectly integrated with Mixed Sample Data Augmentation, so we paste the thumbnail in another image where the ground truth labels are also mixed with a certain weight, which makes great achievements on various computer vision tasks. Extensive experiments show that Thumbnail works better than the state-of-the-art augmentation strategies across classification, fine-grained image classification, and object detection. On ImageNet classification, ResNet50 architecture with our method achieves 79.21% accuracy, which is more than 2.89% improvement on the baseline.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
