An RFI Mitigation Pipeline for CRAFTS Multi-beam Data Based on Signal Cross-Correlation Function and SumThreshold Algorithm

The increasing radio frequency interference(RFI)is a well-recognized problem in radio astronomy research.Pulsars and Fast Radio Bursts(FRBs)are high-priority science targets of the ongoing Commercial Radio Astronomy FAST Survey(CRAFTS).To improve the quality of RFI removal in searches of pulsars and FRBs based on CRAFTS multi-beam data,we here propose an intuitive but powerful RFI mitigation pipeline(CCF-ST).The“CCF-ST”is a spatial filter constructed by signal cross-correlation function(CCF)and Sum-Threshold(ST)algorithm.The RFI marking result is saved in a“mask”file,a binary format for RFI masks in PRESTO.Three known pulsars,PSR B0525-21,PSR B0621-04,and PSR J0943+2252 from CRAFTS L-band 19 beams data are used for evaluation of the performance of CCF-ST in comparison with other methods,such as PRESTO’s“rfifind”,ArPLS-ST and ArPLS-SF.The result shows that CCF-ST can reduce effective data loss rate and improves the detected signal-to-noise ratio of the pulsations by~26%and~18%respectively compared with PRESTO’s“rfifind”and ArPLS-ST.The CCF-ST also has the advantage of low computational cost,e.g.,reducing the time consumption by~40%and memory consumption by~90%compared with ArPLS-SF.We expect that the new RFI mitigation and analysis toolkit(CCF-ST)demonstrated in this paper can be applied to CRAFTS and other multi-beam telescope observations to improve the data quality and efficiency of pulsar and FRB searches.

Updated Inventory of Carbon Monoxide in the Taurus Molecular Cloud

The most extensive survey of carbon monoxide(CO)gas in the Taurus molecular cloud relied on ^(12)CO and ^(13)CO J=1→0 emission only,distinguishing the region where ^(12)CO is detected without ^(13)CO(named mask 1 region)from the one where both are detected(mask 2 region)(Goldsmith et al.2008;Pineda et al.2010).We have taken advantage of recent ^(12)CO J=3→2 James Clerk Maxwell Telescope observations,where they include mask 1regions to estimate density,temperature,and N(CO)with a large velocity gradient model.This represents 1395 pixels out of~1.2 million in the mark 1 region.Compared to Pineda et al.(2010)results and assuming a Tkin of 30 K,we find a higher volume density of molecular hydrogen of 3.3×10^(3) cm^(-3),compared to their 250-700 cm^(-3),and a CO column density of 5.7×10^(15)cm^(-2),about a quarter of their value.The differences are important and show the necessity to observe several CO transitions to better describe the intermediate region between the dense cloud and the diffuse atomic medium.Future observations to extend the ^(12)CO J=3→2 mapping further away from the ^(13)COdetected region comprising mask 1 are needed to revisit our understanding of the diffuse portions of dark clouds.

An Overview of FAST Real-time Fast Radio Burst Searching System

In this paper,we report a real-time Fast Radio Burst(FRB)searching system that has been successfully implemented with the 19 beam receiver of the Five-hundred-meter Aperture Spherical radio Telescope(FAST).The relatively small field of view of FAST makes the search for new FRBs challenging,but its high sensitivity significantly improves the accuracy of FRB localization and enables the detection of high-precision neutral hydrogen absorption lines generated by FRBs.Our goal is to develop an FRB searching system capable of realtime detection of FRBs that allows high-time resolution spectro-temporal studies among the repeated bursts,as well as detailed investigations of these bursts and exploration of FRB progenitor models.The data from each beam of the 19-beam receiver are fed into a high-performance computing node server,which performs real-time searches for pulses with a wide dispersion measure(DM)range of 20–10,000 pc cm^(-3) with step efficiency of 25%in real time.Then,the head node server aggregates all the candidate signals from each beam within a given time,determining their authenticity based on various criteria,including arrival time,pulse width,signal-to-noise ratio and coincidence patterns among the 19 beams.Within the 1.05–1.45 GHz operating bandwidth of the FAST 19beam receiver,the system achieves a frequency resolution of 122.07 kHz and a time resolution of 270.336μs.Subsequently,our team detected a series of bursts with a DM of 566 on 2019 August 30 confirming them as FRB121102.The FRB searching system enables the 19-beam receiver of FAST to detect repeated/one-off pulses/bursts in real time.

Cross-correlation Forecast of CSST Spectroscopic Galaxy and MeerKAT Neutral Hydrogen Intensity Mapping Surveys

Cross-correlating the data on neutral hydrogen(HⅠ)21 cm intensity mapping with galaxy surveys is an effective method to extract astrophysical and cosmological information.In this work,we investigate the cross-correlation of MeerKAT single-dish mode HⅠintensity mapping and China Space Station Telescope(CSST)spectroscopic galaxy surveys.We simulate a survey area of~300 deg~2 of MeerKAT and CSST surveys at z=0.5 using MultiDark N-body simulation.The PC A algorithm is applied to remove the foregrounds of HⅠintensity mapping,and signal compensation is considered to solve the signal loss problem in HⅠ-galaxy cross power spectrum caused by the foreground removal process.We find that from CSST galaxy auto and MeerKAT-CSST cross power spectra,the constraint accuracy of the parameter productΩ_(HⅠ)b_(HⅠ)r_(HⅠ,g)can reach~1%,which is about one order of magnitude higher than the current results.After performing the full MeerKAT HⅠintensity mapping survey with5000 deg~2 survey area,the accuracy can be enhanced to<0.3%.This implies that the MeerKAT-CSST cross-correlation can be a powerful tool to probe the cosmic HⅠproperty and the evolution of galaxies and the Universe.

A GPU based single-pulse search pipeline(GSP)with database and its application to the Commensal Radio Astronomy FAST Survey(CRAFTS)

We developed a GPU based single-pulse search pipeline(GSP)with a candidate-archiving database.Largely based upon the infrastructure of the open source PulsaR Exploration and Search Toolkit(PRESTO),GSP implements GPU acceleration of the de-dispersion and integrates a candidate-archiving database.We applied GSP to the data streams from the Commensal Radio Astronomy FAST Survey(CRAFTS),which resulted in quasi-real-time processing.The integrated candidate database facilitates synergistic usage of multiple machine-learning tools and thus improves efficient identification of radio pulsars such as rotating radio transients(RRATs)and fast radio bursts(FRBs).We first tested GSP on pilot CRAFTS observations with the FAST Ultra-Wide Band(UWB)receiver.GSP detected all pulsars known from the the Parkes multibeam pulsar survey in the corresponding sky area covered by the FAST-UWB.GSP also discovered 13 new pulsars.We measured the computational efficiency of GSP to be~120 times faster than the original PRESTO and~60 times faster than an MPI-parallelized version of PRESTO.

An Arecibo follow-up study of seven pulsars discovered by Five-hundred-meter Aperture Spherical radio Telescope(FAST)

We present Arecibo 327 MHz confirmation and follow-up studies of seven new pulsars discovered by the Five-hundred-meter Aperture Spherical radio Telescope(FAST).These pulsars are discovered in a pilot program of the Commensal Radio Astronomy FAST Survey(CRAFTS)with the ultra-widebandwidth commissioning receiver.Five of them are normal pulsars and two are extreme nulling slow pulsars.PSR J2111+2132’s dispersion measure(DM:78.5 pc cm^(-3))is above the upper limits of the two Galactic free electron density models,NE2001 and YMW16,and PSR J2057+2133’s position is out of the Scutum-Crux Arm,making them uniquely useful for improving the Galactic free electron density model in their directions.We present a detailed single pulse analysis for the slow nulling pulsars.We show evidence that PSR J2323+1214’s main pulse component follows a non-Poisson distribution and marginal evidence for a sub-pulse-drift or recurrent period of 32.3±0.4 rotations from PSR J0539+0013.We discuss the implication of our finding to the pulsar radiation mechanism.

The Potential of Detecting Radio-flaring Ultracool Dwarfs at L band in the FAST Drift-scan Survey

The Five-hundred-meter Aperture Spherical radio Telescope(FAST)completed its commissioning and began the Commensal Radio Astronomy Fas T Survey(CRAFTS),a multi-year survey to cover 60%of the sky,in 2020.We present predictions for the number of radio-flaring ultracool dwarfs(UCDs)that are likely to be detected by CRAFTS.Based on the observed flaring UCDs from a number of unbiased,targeted radio surveys in the literature,we derive a detection rate of≥3%.Assuming a flat radio spectrumνLν∝ν^(β+1) withβ=-1.0 for UCD flares,we construct a flare luminosity function dN/dL∝L^(-1.96±0.45)(here L=νLν).CRAFTS is found to be sensitive enough for flares from UCDs up to~180 pc.Considering the Galactic thin disk,we carry out a 3D Monte Carlo simulation of the UCD population,which is then fed to mock CRAFTS observations.We estimate that~170 flaring UCDs would be detected through transient searches in circular polarization.Though only marginally sensitive to the scale height of UCDs,the results are very sensitive to the assumed spectral indexβ.Forβfrom 0 to-2.5,the number of expected detections increases dramatically from~20 to~3460.We also contemplate the strategies for following up candidates of flaring UCDs,and discuss the implications of survey results for improving our knowledge of UCD behavior at L band and dynamos.

Collapsing index:a new method to identify star-forming cores based on ALMA images

Stars form through the gravitational collapse of molecular cloud cores.Before collapsing,the cores are supported by thermal pressure and turbulent motions.A question of critical importance for the understanding of star formation is how to observationally discern whether a core has already initiated gravitational collapse or is still in hydrostatic balance.The canonical method to identify gravitational collapse is based on the observed radial density profile,which would change from Bonnor-Ebert type toward power laws as the core collapses.In practice,due to the projection effect,the resolution limit and other caveats,it has been difficult to directly reveal the dynamical status of cores,particularly in massive star forming regions.We here propose a novel,straightforward diagnostic,namely,the collapsing index(CI),which can be modeled and calculated based on the radial profile of the line width of dense gas.A meaningful measurement of CI requires spatially and spectrally resolved images of optically thin and chemically stable dense gas tracers.ALMA observations are making such data sets increasingly available for massive star forming regions.Applying our method to one of the deepest dense-gas spectral images ever taken toward such a region,namely,the Orion molecular cloud,we detect the dynamical status of selected cores.We observationally distinguished a collapsing core in a massive star forming region from a hydrostatical one.Our approach would help significantly improve our understanding of the interaction between gravity and turbulence within molecular cloud cores in the process of star formation.

Simulation of the gravitational wave frequency distribution of neutron star-black hole mergers

LIGO-Virgo has observed the gravitational waves(GWs)from the coalescence of binary black hole(BBH)and binary neutron star(BNS)during O1 and O2,and the ones from NS-BH are expected to be hunted in the operating O3 run.The population properties and mass distribution of NS-BH mergers are poorly understood now,thus researchers simulated their chirp mass(M)distribution by a synthetic model,in which the BHs and NSs were inferred by LIGO-Virgo(O1/O2),and obtained the values in the range of 2.1 M_(⊙)<M<7.3 M_(⊙).In this paper,we further simulate the GW frequency(fGW)distribution of NS-BH mergers by the above-stated synthetic model,with a basic binary system model through the Monte Carlo method.Our results predict that the median with 90%credible intervals is 165+475-64 Hz in the case of Schwarzschild BH when the system just before merger,and this GW frequency is expected to increase several times in the merger stage,which is lying in the frequency band of LIGO-Virgo,i.e.,about 15 Hz to a few kHz.Our results provide an important reference for hunting the NS-BH mergers by the on-going O3 run of LIGO-Virgo.

Interacting system NGC 7805/6(Arp 112)and its tidal dwarf galaxy candidate

We present results from our Giant Metrewave Radio Telescope(GMRT)HⅠ,Himalayan Chandra Telescope(HCT)Hα,1 m Sampurnanand Telescope(ST)and 1.3 m Devasthal Fast Optical Telescope(DFOT)deep optical observations of the NGC 7805/6(Arp 112)system to test KUG 2359+311’s tidal dwarf galaxy(TDG)candidacy and explore the properties of the interacting system.Our GMRT HⅠmap shows no HⅠdetection associated with KUG 2359+311,nor any HⅠtail or bridge-like structure connecting KUG 2359+311 to the NGC 7805/6 system.Our HCT Hαimage,on the other hand,displays strong detections in KUG 2359+311,with net SFR0.035±0.009 M_(⊙)yr^(-1).The Hαdata constrain the redshift of KUG 2359+311 to 0.00≤z≤0.043,compared to the redshift of NGC 7806 of0.015.TDGs detected to date have all been HⅠrich,and displayed HⅠ,ionised gas and stellar tidal debris trails(bridges or tails)linking them to their parent systems.However,neither our HⅠdata nor our optical imaging,while three magnitudes deeper than SDSS,reveals a tidal trail connecting KUG 2359+311 to NGC 7805/6.Lack of HⅠ,presence of an old stellar population,ongoing star formation and reasonably high SFR compared to normal dwarf galaxies suggest that KUG 2359+311 may not be an Arp 112 TDG.It is most likely a case of a regular gas-rich dwarf galaxy undergoing a morphological transformation after having lost its entire gas content to an interaction with the Arp 112 system.Redshift and metallicity from future spectroscopic observations of KUG 2359+311 would help clarify the nature of this enigmatic structure.

Resolution-dependent subsonic non-thermal line dispersion revealed by ALMA

We report here Atacama Large Millimeter/submillimeter Array(ALMA)N2H+(1-0)images of the Orion Molecular Cloud 2 and 3(OMC-2/3)with high angular resolution(3"or 1200 au)and high spatial dynamic range.Combining a dataset from the ALMA main array,Atacama Compact Array(ACA),Nobeyama 45-m Telescope and Very Large Array(VLA)(providing temperature measurement on matching scales),we find that most of the dense gas in OMC-2/3 is subsonic(σQNT/cs=0.62)with a mean line width(△v)of 0.39 kms-1 full width at half maximum(FWHM).This is markedly different from the majority of previous observations of massive star-forming regions.In contrast,line widths from the Nobeyama Telescope are transonic at 0.69 km s-1(σNT/cs=1.08).We demonstrated that the larger line widths obtained by the single-dish telescope arose from unresolved sub-structures within their respective beams.The dispersions from larger scalesσls(as traced by the Nobeyama Telescope)can be decomposed into three components such thatσls2=σss2+σbm2+σrd2,where small-scaleσss is the line dispersion of each ALMA beam,bulk motionσbm is dispersion between peak velocity of each ALMA beam andσrd is the residual dispersion.Such decomposition,though purely empirical,appears to be robust throughout our data cubes.Apparent supersonic line widths,commonly found in massive molecular clouds,are thus likely due to the effect of poor spatial resolution.The observed non-thermal line dispersion(sometimes referred to as’turbulence’)transits from supersonic to subsonic at~0.05 pc scales in the OMC-2/3 region.Such transition could be commonly found with sufficient spatial(not just angular)resolution,even in regions with massive young clusters,such as the Orion molecular clouds studied here.

Studying infall in infrared dark clouds with multiple HCO transitions

We investigate the infall properties in a sample of 11 infrared dark clouds(IRDCs) showing blue-asymmetry signatures in HCO^(+)J=1-0 line profiles.We used JCMT to conduct mapping observations in HCO^(+)J=4-3 as well as single-point observations in HCO+J=3-2,towards 23 clumps in these IRDCs.We applied the HILL model to fit these observations and derived infall velocities in the range of 0.5-2.7 km s^(-1),with a median value of 1.0 km s^(-1),and obtained mass accretion rates of 0.5-14 ×10^(-3) Mo yr^(-1).These values are comparable to those found in massive star forming clumps in later evolutionary stages.These IRDC clumps are more likely to form star clusters.HCO^(+)J=3-2 and HCO^(+)J=1-0 were shown to trace infall signatures well in these IRDCs with comparable inferred properties.HCO^(+)J=4-3,on the other hand,exhibits infall signatures only in a few very massive clumps,due to smaller opacities.No obvious correlation for these clumps was found between infall velocity and the NH3/CCS ratio.

Radio frequency interference mitigation using pseudoinverse learning autoencoders

Radio frequency interference(RFI)is an important challenge in radio astronomy.RFI comes from various sources and increasingly impacts astronomical observation as telescopes become more sensitive.In this study,we propose a fast and effective method for removing RFI in pulsar data.We use pseudo-inverse learning to train a single hidden layer auto-encoder(AE).We demonstrate that the AE can quickly learn the RFI signatures and then remove them from fast-sampled spectra,leaving real pulsar signals.This method has the advantage over traditional threshold-based filter method in that it does not completely remove contaminated channels,which could also contain useful astronomical information.

FAST Observations of an Extremely Active Episode of FRB 20201124A.Ⅱ.Energy Distribution

We report the properties of more than 800 bursts detected from the repeating fast radio burst(FRB)source FRB20201124A with the Five-hundred-meter Aperture Spherical radio Telescope(FAST)during an extremely active episode on UTC 2021 September 25–28 in a series of four papers.In this second paper of the series,we study the energy distribution of 881 bursts(defined as significant signals separated by dips down to the noise level)detected in the first four days of our 19 hr observational campaign spanning 17 days.The event rate initially increased exponentially but the source activity stopped within 24 hr after the 4th day.The detection of 542 bursts in one hour during the fourth day marked the highest event rate detected from one single FRB source so far.The bursts have complex structures in the time-frequency space.We find a double-peak distribution of the waiting time,which can be modeled with two log-normal functions peaking at 51.22 ms and 10.05 s,respectively.Compared with the emission from a previous active episode of the source detected with FAST,the second distribution peak time is smaller,suggesting that this peak is defined by the activity level of the source.We calculate the isotropic energy of the bursts using both a partial bandwidth and a full bandwidth and find that the energy distribution is not significantly changed.We find that an exponentially connected broken-power law function can fit the cumulative burst energy distribution well,with the lower and higher-energy indices being-1.22±0.01 and-4.27±0.23,respectively.Assuming a radio radiative efficiency ofη_(r)=10^(-4),the total isotropic energy of the bursts released during the four days when the source was active is already 3.9×10^(46)erg,exceeding~23%of the available magnetar dipolar magnetic energy.This challenges the magnetar models which invoke an inefficient radio emission(e.g.,synchrotron maser models).

Revisiting pulsar velocities using Gaia Data Release 2

Precise measurements of neutron star(NS)velocities provide critical clues in regard to the supernova physics and evolution of binary systems.Based on Gaia Data Release 2(DR2),we selected a sample of 24 young(<3 Myr)pulsars with precise parallax measurements and measured the velocity of their local standard of rest(LSR)and the velocity dispersion among their respective local stellar groups.The median velocity difference between thus calculated LSRs and the Galactic rotation model is~7.6 km s^(-1),small compared to the typical velocity dispersion of~27.5 km s^(-1).For pulsars off the Galactic plane,such differences grow significantly to as large as~40 km s^(-1).More importantly,the velocity dispersion of stars in the local group of low-velocity pulsars can be comparable to their transverse velocities,suggesting that the intrinsic velocities of NS progenitors should be taken into account when we consider their natal kicks and binary evolution.We also examined the double NS system J0737-3039 A/B,and measured its transverse velocity to be 26-13+18 km s^(-1)assuming nearby Gaia sources are representative of their birth environment.This work demonstrated the feasibility and importance of using Gaia data to study the velocity of individual systems and velocity distribution of NSs.

Statistical properties of fast radio bursts elucidate their origins:magnetars are favored over gamma-ray bursts

Fast radio bursts(FRBs) are extremely strong radio flares lasting several milliseconds,most of which come from unidentified objects at a cosmological distance.They can be apparently repeating or not.In this paper,we analyzed 18 repeaters and 12 non-repeating FRBs observed in the frequency bands of 400–800 MHz from Canadian Hydrogen Intensity Mapping Experiment(CHIME).We investigated the distributions of FRB isotropic-equivalent radio luminosity,considering the K correction.Statistically,the luminosity distribution can be better fitted by Gaussian form than by power-law.Based on the above results,together with the observed FRB event rate,pulse duration,and radio luminosity,FRB origin models are evaluated and constrained such that the gamma-ray bursts(GRBs) may be excluded for the non-repeaters while magnetars or neutron stars(NSs) emitting the supergiant pulses are preferred for the repeaters.We also found the necessity of a small FRB emission beaming solid angle(about 0.1 sr) from magnetars that should be considered,and/or the FRB association with soft gamma-ray repeaters(SGRs) may lie at a low probability of about 10%.Finally,we discussed the uncertainty of FRB luminosity caused by the estimation of the distance that is inferred by the simple relation between the redshift and dispersion measure(DM).

FAST Observations of an Extremely Active Episode of FRB 20201124A.Ⅲ.Polarimetry

As the third paper in the multiple-part series,we report the statistical properties of radio bursts detected from the repeating fast radio burst(FRB)source FRB 20201124A with the Five-hundred-meter Aperture Spherical radio Telescope during an extremely active episode between the 25th and 28th of September 2021(UT).We focus on the polarization properties of536 bright bursts with S/N>50.We found that the Faraday rotation measures(RMs)monotonically dropped from-579to-605 rad m^(-2)in the 4 day window.The RM values were compatible with the values(-300 to-900 rad m^(-2))reported 4 months ago.However,the RM evolution rate in the current observation window was at least an order of magnitude smaller than the one(~500 rad m^(-2)day^(-1))previously reported during the rapid RM-variation phase,but is still higher than the one(≤1 rad m^(-2)day^(-1))during the later RM no-evolution phase.The bursts of FRB 20201124A were highly polarized with the total degree of polarization(circular plus linear)greater than 90%for more than 90%of all bursts.The distribution of linear polarization position angles(PAs),degree of linear polarization(L/I)and degree of circular polarization(V/I)can be characterized with unimodal distribution functions.During the observation window,the distributions became wider with time,i.e.,with larger scatter,but the centroids of the distribution functions remained nearly constant.For individual bursts,significant PA variations(confidence level 5σ)were observed in 33%of all bursts.The polarization of single pulses seems to follow certain complex trajectories on the Poincarésphere,which may shed light on the radiation mechanism at the source or the plasma properties along the path of FRB propagation.

FAST Observations of an Extremely Active Episode of FRB 20201124A.Ⅳ.Spin Period Search

We report the properties of more than 800 bursts detected from the repeating fast radio burst(FRB)source FRB 20201124A with the Five-hundred-meter Aperture Spherical radio Telescope during an extremely active episode on UTC 2021 September 25th-28th in a series of four papers.In this fourth paper of the series,we present a systematic search of the spin period and linear acceleration of the source object from both 996 individual pulse peaks and the dedispersed time series.No credible spin period was found from this data set.We rule out the presence of significant periodicity in the range between 1 ms and 100 s with a pulse duty cycle<0.49±0.08(when the profile is defined by a von-Mises function,not a boxcar function)and linear acceleration up to 300 m s^(-2)in each of the four one-hour observing sessions,and up to 0.6 m s^(-2)in all 4 days.These searches contest theoretical scenarios involving a 1 ms–100 s isolated magnetar/pulsar with surface magnetic field<10^(15)G and a small duty cycle(such as in a polar-cap emission mode)or a pulsar with a companion star or black hole up to 100 M_(⊙)and P_(b)>10 hr.We also perform a periodicity search of the fine structures and identify 53 unrelated millisecond-timescale“periods”in multicomponents with the highest significance of 3.9σ.The“periods”recovered from the fine structures are neither consistent nor harmonically related.Thus they are not likely to come from a spin period.We caution against claiming spin periodicity with significance below~4σwith multi-components from one-off FRBs.We discuss the implications of our results and the possible connections between FRB multi-components and pulsar microstructures.

Scintillation Arc from FRB 20220912A

We present the interstellar scintillation analysis of fast radio burst(FRB)20220912A during its extremely active episode in 2022using data from the Five-hundred-meter Aperture Spherical Radio Telescope(FAST).We detect a scintillation arc in the FRB’s secondary spectrum,which describes the power in terms of the scattered FRB signals’time delay and Doppler shift.The arc indicates that the scintillation is caused by a highly localized region.Our analysis favors a Milky Way origin of the ionized interstellar medium(IISM)for the localized scattering medium but cannot rule out a host galaxy origin.We present our method for detecting the scintillation arc,which can be applied generally to sources with irregularly spaced bursts or pulses.These methods could help shed light on the complex interstellar environment surrounding the FRBs and in our Galaxy.

Machine learning for nanohertz gravitational wave detection and parameter estimation with pulsar timing array

Studies have shown that the use of pulsar timing arrays(PTAs)is among the approaches with the highest potential to detect very low-frequency gravitational waves in the near future.Although the capture of gravitational waves(GWs)by PTAs has not been reported yet,many related theoretical studies and some meaningful detection limits have been reported.In this study,we focused on the nanohertz GWs from individual supermassive binary black holes.Given specific pulsars(PSR J1909-3744,PSR J1713+0747,PSR J0437-4715),the corresponding GW-induced timing residuals in PTAs with Gaussian white noise can be simulated.Further,we report the classification of the simulated PTA data and parameter estimation for potential GW sources using machine learning based on neural networks.As a classifier,the convolutional neural network shows high accuracy when the combined signal to noise ratio≥1.33 for our simulated data.Further,we applied a recurrent neural network to estimate the chirp mass(M)of the source and luminosity distance(Dp)of the pulsars and Bayesian neural networks(BNNs)to obtain the uncertainties of chirp mass estimation.Knowledge of the uncertainties is crucial to astrophysical observation.In our case,the mean relative error of chirp mass estimation is less than 13.6%.Although these results are achieved for simulated PTA data,we believe that they will be important for realizing intelligent processing in PTA data analysis.