The paper reports that SHAP-identified top features of deep learning models trained on BRFSS 2023 align with medical literature for 13 chronic diseases, and uses this alignment to argue the models are broadly trustworthy.
The Galactic population and properties of young, highly-energetic pulsars
1 Pith paper cite this work. Polarity classification is still indexing.
abstract
The population of young, non-recycled pulsars with spin down energies Edot >10^35 erg/s is sampled predominantly at gamma-ray and radio wavelengths. A total of 137 such pulsars are known, with partial overlap between the sources detectable in radio and gamma-rays. We use a very small set of assumptions in an attempt to test whether the observed pulsar sample can be explained by a single underlying population of neutron stars. For radio emission we assume a canonical conal beam with a fixed emission height of 300~km across all spin periods and a luminosity law which depends on Edot^{0.25}. For gamma-ray emission we assume the outer-gap model and a luminosity law which depends on Edot^{0.5}. We synthesise a population of fast-spinning pulsars with a birth rate of one per 100 years. We find that this simple model can reproduce most characteristics of the observed population with two caveats. The first is a deficit of gamma-ray pulsars at the highest Edot which we surmise to be an observational selection effect due to the difficulties of finding gamma-ray pulsars in the presence of glitches without prior knowledge from radio frequencies. The second is a deficit of radio pulsars with interpulse emission, which may be related to radio emission physics. We discuss the implications of these findings.
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Trustworthy Chronic Disease Risk Prediction For Self-Directed Preventive Care via Medical Literature Validation
The paper reports that SHAP-identified top features of deep learning models trained on BRFSS 2023 align with medical literature for 13 chronic diseases, and uses this alignment to argue the models are broadly trustworthy.