In the third catalog of active galactic nuclei detected by the $Fermi$ Large Area Telescope Clean (3LAC) sample, there are 402 blazars candidates of uncertain type (BCU). The proposed analysis will help to evaluate the potential optical classification flat spectrum radio quasars (FSRQs) versus BL Lacertae (BL Lacs) objects of BCUs, which can help to understand which is the most elusive class of blazar hidden in the Fermi sample. By studying the 3LAC sample, we found some critical values of $\gamma$-ray photon spectral index ($\Gamma_{\rm ph}$), variability index (VI) and radio flux (${\rm F_R}$) of the sources separate known FSRQs and BL Lac objects. We further utilize those values to defined an empirical "high-confidence" candidate zone that can be used to classify the BCUs. Within such a zone ($\Gamma_{\rm ph}<2.187$, log${\rm F_R}<2.258$ and ${ \rm logVI <1.702}$), we found that 120 BCUs can be classified BL Lac candidates with a higher degree of confidence (with a misjudged rate $<1\%$). Our results suggest that an empirical "high confidence" diagnosis is possible to distinguish the BL Lacs from the Fermi observations based on only on the direct observational data of $\Gamma_{\rm ph}$, VI and ${\rm F_R}$.

Machine learning allows efficient extraction of physical properties from stellar spectra that have been obtained by large surveys. The viability of ML approaches has been demonstrated for spectra covering a variety of wavelengths and spectral resolutions, but most often for main sequence or evolved stars, where reliable synthetic spectra provide labels and data for training. Spectral models of young stellar objects (YSOs) and low mass main sequence (MS) stars are less well-matched to their empirical counterparts, however, posing barriers to previous approaches to classify spectra of such stars. In this work we generate labels for YSOs and low mass MS stars through their photometry. We then use these labels to train a deep convolutional neural network to predict log g, Teff, and Fe/H for stars with APOGEE spectra in the DR14 dataset. This "APOGEE Net" has produced reliable predictions of log g for YSOs, with uncertainties of within 0.1 dex and a good agreement with the structure indicated by pre-main sequence evolutionary tracks, and correlate well with independently derived stellar radii. These values will be useful for studying pre-main sequence stellar populations to accurately diagnose membership and ages.

The recently published fourth Fermi Large Area Telescope source catalog (4FGL) reports 5065 gamma-ray sources in terms of direct observational gamma-ray properties. Among the sources, the largest population is the Active Galactic Nuclei (AGN), which consists of 3137 blazars, 42 radio galaxies, and 28 other AGNs. The blazar sample comprises 694 flat-spectrum radio quasars (FSRQs), 1131 BL Lac-type objects (BL Lacs), and 1312 blazar candidates of an unknown type (BCUs). The classification of blazars is difficult using optical spectroscopy given the limited knowledge with respect to their intrinsic properties, and the limited availability of astronomical observations. To overcome these challenges, machine learning algorithms are being investigated as alternative approaches. Using the 4FGL catalog, a sample of 3137 Fermi blazars with 23 parameters is systematically selected. Three established supervised machine learning algorithms (random forests (RFs), support vector machines (SVMs), artificial neural networks (ANNs)) are employed to general predictive models to classify the BCUs. We analyze the results for all of the different combinations of parameters. Interestingly, a previously reported trend the use of more parameters leading to higher accuracy is not found. Considering the least number of parameters used, combinations of eight, 12 or 10 parameters in the SVM, ANN, or RF generated models achieve the highest accuracy (Accuracy $\simeq$ 91.8\%, or $\simeq$ 92.9\%). Using the combined classification results from the optimal combinations of parameters, 724 BL Lac type candidates and 332 FSRQ type candidates are predicted; however, 256 remain without a clear prediction.

In the third catalog of active galactic nuclei detected by the Fermi-LAT (3LAC) Clean Sample, there are 402 blazars candidates of uncertain type (BCU). Due to the limitations of astronomical observation or intrinsic properties, it is difficult to classify blazars using optical spectroscopy. The potential classification of BCUs using machine learning algorithms is essential. Based on the 3LAC Clean Sample, we collect 1420 Fermi blazars with 8 parameters of {\gamma}-ray photon spectral index, radio flux, flux density, curve significance, the integral photon flux in 100 to 300 MeV, 0.3 to 1 GeV, 10 to 100 GeV and variability index. Here, we apply 4 different supervised machine learning (SML) algorithms (\emph{Decision trees, Random forests, support vector machines and Mclust Gaussian finite mixture models}) to evaluate the classification of BCUs based on the direct observational properties. All the 4 methods can perform exceedingly well with a more accuracy and can effective forecast the classification of Fermi BCUs. The evaluating results show the results of these methods (SML) are valid and robust, where, about 1/4 sources are FSRQs and 3/4 are BL Lacs in 400 BCUs, which are consistent with some other recent results. Although a number of factors influence the accuracy of SML, the results are stable at a fixed ratio 1:3 between FSRQs and BL Lacs, which suggests that the SML can provides an effective method to evaluate the potential classification of BCUs. Among the 4 methods, Mclust Gaussian Mixture Modelling has the highest accuracy for our training sample (4/5, seed=123).

The curvature of the $\gamma$-ray spectrum in blazars may reflect the intrinsic distribution of the emitting electron distribution, which will further give some information on the possible acceleration and cooling processes in the emitting region. The $\gamma$-ray spectra of Fermi blazars are normally fitted either by a single power-law (PL) or a log-normal (call Logarithmic Parabola, LP) form. The possible reason for this differnece is not unclear. We statistically explore this issue based on the different observational properties of 1419 Fermi blazars in the 3LAC Clean sample. We find that the $\gamma$-ray flux (100 MeV-100 GeV) and variability index follow bimodal distributions for PL and LP blazars, where $\gamma$-ray flux and variability index show {a positive correlation}. However, the distributions of the $\gamma$-ray luminosity and redshift follow a unimodal distribution. Our results suggest that the bimodal distribution of $\gamma$-ray flux for LP and PL blazars may be not intrinsic and all blazars may have an intrinsic curved $\gamma$-ray spectrum and the PL spectrum is just caused by the fitting effect due to the less photons.

Previous studies have found that the width of gamma-ray burst (GRB) pulse is energy dependent and that it decreases as a power-law function with increasing photon energy. In this work we have investigated the relation between the energy dependence of pulse and the so-called Band spectrum by using a sample including 51 well-separated fast rise and exponential decay long-duration GRB pulses observed by BATSE (Burst and Transient Source Experiment on the Compton Gamma Ray Observatory). We first decompose these pulses into rise, and decay phases and find the rise widths, and the decay widths also behavior as a power-law function with photon energy. Then we investigate statistically the relations between the three power-law indices of the rise, decay and total width of pulse (denoted as $\delta_r$, $\delta_d$ and $\delta_w$, respectively) and the three Band spectral parameters, high-energy index ($\alpha$), low-energy index ($\beta$) and peak energy ($E_p$). It is found that (1)$\alpha$ is strongly correlated with $\delta_w$ and $\delta_d$ but seems uncorrelated with $\delta_r$; (2)$\beta$ is weakly correlated with the three power-law indices and (3)$E_p$ does not show evident correlations with the three power-law indices. We further investigate the origin of $\delta_d-\alpha$ and $\delta_w-\alpha$. We show that the curvature effect and the intrinsic Band spectrum could naturally lead to the energy dependence of GRB pulse width and also the $\delta_d-\alpha$ and $\delta_w-\alpha$ correlations. Our results would hold so long as the shell emitting gamma rays has a curve surface and the intrinsic spectrum is a Band spectrum or broken power law. The strong $\delta_d-\alpha$ correlation and inapparent correlations between $\delta_r$ and three Band spectral parameters also suggest that the rise and decay phases of GRB pulses have different origins.

CdZnTe (CZT) is a wide bandgap II-VI semiconductor developed for the spectroscopic detection of X-rays and {\gamma}-rays at room temperature. The Swift Burst Alert Telescope is using an 5240 cm2 array of 2 mm thick CZT detectors for the detection of 15-150 keV X-rays from Gamma-Ray Bursts. We report on the systematic tests of thicker (\geq 0.5 cm) CZT detectors with volumes between 2 cm3 and 4 cm3 which are potential detector choices for a number of future X-ray telescopes that operate in the 10 keV to a few MeV energy range. The detectors contacted in our laboratory achieve Full Width Half Maximum energy resolutions of 2.7 keV (4.5%) at 59 keV, 3 keV (2.5%) at 122 keV and 4 keV (0.6%) at 662 keV. The 59 keV and 122 keV energy resolutions are among the world-best results for \geq 0.5 cm thick CZT detectors. We use the data set to study trends of how the energy resolution depends on the detector thickness and on the pixel pitch. Unfortunately, we do not find clear trends, indicating that even for the extremely good energy resolutions reported here, the achievable energy resolutions are largely determined by the properties of individual crystals. Somewhat surprisingly, we achieve the reported results without applying a correction of the anode signals for the depth of the interaction. Measuring the interaction depths thus does not seem to be a pre-requisite for achieving sub-1% energy resolutions at 662 keV.

Assuming an intrinsic `Band' shape spectrum and an intrinsic energy-independent emission profile we have investigated the connection between the evolution of the rest-frame spectral parameters and the spectral lags measured in gamma-ray burst (GRB) pulses by using a pulse model. We first focus our attention on the evolution of the peak energy, $E_{0,p}$, and neglect the effect of the curvature effect. It is found that the evolution of $E_{0,p}$ alone can produce the observed lags. When $E_{0,p}$ varies from hard to soft only the positive lags can be observed. The negative lags would occur in the case of $E_{0,p}$ varying from soft to hard. When the evolution of $E_{0,p}$ and the low-energy spectral index $\alpha_{0}$ varying from soft to hard then to soft we can find the aforesaid two sorts of lags. We then examine the combined case of the spectral evolution and the curvature effect of fireball and find the observed spectral lags would increase. A sample including 15 single pulses whose spectral evolution follows hard to soft has been investigated. All the lags of these pulses are positive, which is in good agreement with our theoretical predictions. Our analysis shows that only the intrinsic spectral evolution can produce the spectral lags and the observed lags should be contributed by the intrinsic spectral evolution and the curvature effect. But it is still unclear what cause the spectral evolution.

In this paper we have analyzed the temporal and spectral behavior of 52 Fast Rise and Exponential Decay (FRED) pulses in 48 long-duration gamma-ray bursts (GRBs) observed by the CGRO/BATSE, using a pulse model with two shape parameters and the Band model with three shape parameters, respectively. It is found that these FRED pulses are distinguished both temporally and spectrally from those in long-lag pulses. Different from these long-lag pulses only one parameter pair indicates an evident correlation among the five parameters, which suggests that at least $\sim$4 parameters are needed to model burst temporal and spectral behavior. In addition, our studies reveal that these FRED pulses have correlated properties: (i) long-duration pulses have harder spectra and are less luminous than short-duration pulses; (ii) the more asymmetric the pulses are the steeper the evolutionary curves of the peak energy ($E_{p}$) in the $\nu f_{\nu}$ spectrum within pulse decay phase are. Our statistical results give some constrains on the current GRB models.

Employing two samples containing of 56 and 59 well-separated FRED (fast rise and exponential decay) gamma-ray burst (GRB) pulses whose spectra are fitted by the Band spectrum and Compton model, respectively, we have investigated the evolutionary slope of $E_{p}$ (where $E_{p}$ is the peak energy in the $\nu F\nu$ spectrum) with time during the pulse decay phase. The bursts in the samples were observed by the Burst and Transient Source Experiment (BATSE) on the Compton Gamma-Ray Observatory. We first test the $E_{p}$ evolutionary slope during the pulse decay phase predicted by Lu et al. (2007) based on the model of highly symmetric expanding fireballs in which the curvature effect of the expanding fireball surface is the key factor concerned. It is found that the evolutionary slopes are normally distributed for both samples and concentrated around the values of 0.73 and 0.76 for Band and Compton model, respectively, which is in good agreement with the theoretical expectation of Lu et al. (2007). However, the inconsistence with their results is that the intrinsic spectra of most of bursts may bear the Comptonized or thermal synchrotron spectrum, rather than the Band spectrum. The relationships between the evolutionary slope and the spectral parameters are also checked. We show the slope is correlated with $E_{p}$ of time-integrated spectra as well as the photon flux but anticorrelated with the lower energy index $\alpha$. In addition, a correlation between the slope and the intrinsic $E_{p}$ derived by using the pseudo-redshift is also identified. The mechanisms of these correlations are unclear currently and the theoretical interpretations are required.

Employing a sample presented by Kaneko et al. (2006) and Kocevski et al. (2003), we select 42 individual tracking pulses (here we defined tracking as the cases in which the hardness follows the same pattern as the flux or count rate time profile) within 36 Gamma-ray Bursts (GRBs) containing 527 time-resolved spectra and investigate the spectral hardness, $E_{peak}$ (where $E_{peak}$ is the maximum of the $\nu F_{\nu}$ spectrum), evolutionary characteristics. The evolution of these pulses follow soft-to-hard-to-soft (the phase of soft-to-hard and hard-to-soft are denoted by rise phase and decay phase, respectively) with time. It is found that the overall characteristics of $E_{peak}$ of our selected sample are: 1) the $E_{peak}$ evolution in the rise phase always start on the high state (the values of $E_{peak}$ are always higher than 50 keV); 2) the spectra of rise phase clearly start at higher energy (the median of $E_{peak}$ are about 300 keV), whereas the spectra of decay phase end at much lower energy (the median of $E_{peak}$ are about 200 keV); 3) the spectra of rise phase are harder than that of the decay phase and the duration of rise phase are much shorter than that of decay phase as well. In other words, for a complete pulse the initial $E_{peak}$ is higher than the final $E_{peak}$ and the duration of initial phase (rise phase) are much shorter than the final phase (decay phase). This results are in good agreement with the predictions of Lu et al. (2007) and current popular view on the production of GRBs. We argue that the spectral evolution of tracking pulses may be relate to both of kinematic and dynamic process even if we currently can not provide further evidences to distinguish which one is dominant. Moreover, our statistical results give some witnesses to constrain the current GRB model.