Engineering topics

Martinez, J. G.

Publications and source records attributed to Martinez, J. G..

3 records

DOE OSTI2016

NEW DISCOVERIES FROM THE ARECIBO 327 MHz DRIFT PULSAR SURVEY RADIO TRANSIENT SEARCH

We present Clusterrank, a new algorithm for identifying dispersed astrophysical pulses. Such pulses are commonly detected from Galactic pulsars and rotating radio transients (RRATs), which are neutron stars with sporadic radio emission. More recently, isolated, highly dispersed pulses dubbed fast radio bursts (FRBs) have been identified as the potential signature of an extragalactic cataclysmic radio source distinct from pulsars and RRATs. Clusterrank helped us discover 14 pulsars and 8 RRATs in data from the Arecibo 327 MHz Drift Pulsar Survey (AO327). The new RRATs have DMs in the range 23.5–86.6 pc cm{sup −3} and periods in the range 0.172–3.901 s. The new pulsars have DMs in the range 23.6–133.3 pc cm{sup −3} and periods in the range 1.249–5.012 s, and include two nullers and a mode-switching object. We estimate an upper limit on the all-sky FRB rate of 10{sup 5} day{sup −1} for bursts with a width of 10 ms and flux density ≳83 mJy. The DMs of all new discoveries are consistent with a Galactic origin. In comparing statistics of the new RRATs with sources from the RRATalog, we find that both sets are drawn from the same period distribution. In contrast, we find that the period distribution of the new pulsars is different from the period distributions of canonical pulsars in the ATNF catalog or pulsars found in AO327 data by a periodicity search. This indicates that Clusterrank is a powerful complement to periodicity searches and uncovers a subset of the pulsar population that has so far been underrepresented in survey results and therefore in Galactic pulsar population models.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2015

Pulsar J0453+1559: A double neutron star system with a large mass asymmetry

To understand the nature of supernovae and neutron star (NS) formation, as well as binary stellar evolution and their interactions, it is important to probe the distribution of NS masses. Until now, all double NS (DNS) systems have been measured as having a mass ratio close to unity (q ≥ 0.91). Here, we report the measurement of the individual masses of the 4.07-day binary pulsar J0453+1559 from measurements of the rate of advance of periastron and Shapiro delay: the mass of the pulsar is M{sub p} = 1.559 ± 0.005 M{sub ⊙} and that of its companion is M{sub c}=1.174±0.004 M{sub ⊙}; q = 0.75. If this companion is also an NS, as indicated by the orbital eccentricity of the system (e = 0.11), then its mass is the smallest precisely measured for any such object. The pulsar has a spin period of 45.7 ms and a spin period derivative of P-dot = (1.8616±0.0007)×10{sup −19} s s{sup −1}; from these, we derive a characteristic age of ∼ 4.1×10{sup 9} years and a magnetic field of ∼ 2.9×10{sup 9} G, i.e., this pulsar was mildly recycled by the accretion of matter from the progenitor of the companion star. This suggests that it was formed with (very approximately) its current mass. Thus, NSs form with a wide range of masses, which is important for understanding their formation in supernovae. It is also important for the search for gravitational waves released during an NS–NS merger: it is now evident that we should not assume that all DNS systems are symmetric.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2014

Searching for pulsars using image pattern recognition

In the modern era of big data, many fields of astronomy are generating huge volumes of data, the analysis of which can sometimes be the limiting factor in research. Fortunately, computer scientists have developed powerful data-mining techniques that can be applied to various fields. In this paper, we present a novel artificial intelligence (AI) program that identifies pulsars from recent surveys by using image pattern recognition with deep neural nets—the PICS (Pulsar Image-based Classification System) AI. The AI mimics human experts and distinguishes pulsars from noise and interference by looking for patterns from candidate plots. Different from other pulsar selection programs that search for expected patterns, the PICS AI is taught the salient features of different pulsars from a set of human-labeled candidates through machine learning. The training candidates are collected from the Pulsar Arecibo L-band Feed Array (PALFA) survey. The information from each pulsar candidate is synthesized in four diagnostic plots, which consist of image data with up to thousands of pixels. The AI takes these data from each candidate as its input and uses thousands of such candidates to train its ∼9000 neurons. The deep neural networks in this AI system grant it superior ability to recognize various types of pulsars as well as their harmonic signals. The trained AI's performance has been validated with a large set of candidates from a different pulsar survey, the Green Bank North Celestial Cap survey. In this completely independent test, the PICS ranked 264 out of 277 pulsar-related candidates, including all 56 previously known pulsars and 208 of their harmonics, in the top 961 (1%) of 90,008 test candidates, missing only 13 harmonics. The first non-pulsar candidate appears at rank 187, following 45 pulsars and 141 harmonics. In other words, 100% of the pulsars were ranked in the top 1% of all candidates, while 80% were ranked higher than any noise or interference. The performance of this system can be improved over time as more training data are accumulated. This AI system has been integrated into the PALFA survey pipeline and has discovered six new pulsars to date.

79 ASTRONOMY AND ASTROPHYSICS↗