In this work, we present a new, high performance algorithm for background rejection in imaging atmospheric Cherenkov telescopes. We build on the already popular machine-learning techniques used in gamma-ray astronomy by the application of the latest
techniques in machine learning, namely recurrent and convolutional neural networks, to the background rejection problem. Use of these machine-learning techniques addresses some of the key challenges encountered in the currently implemented algorithms and helps to significantly increase the background rejection performance at all energies. We apply these machine learning techniques to the H.E.S.S. telescope array, first testing their performance on simulated data and then applying the analysis to two well known gamma-ray sources. With real observational data we find significantly improved performance over the current standard methods, with a 20-25% reduction in the background rate when applying the recurrent neural network analysis. Importantly, we also find that the convolutional neural network results are strongly dependent on the sky brightness in the source region which has important implications for the future implementation of this method in Cherenkov telescope analysis.
eta Car is the only colliding-wind binary for which high-energy gamma rays are detected. Although the physical conditions in the shock region change on timescales of hours to days, the variability seen at GeV energies is weak and on significantly lon
ger timescales. The gamma-ray spectrum exhibits two features that can be interpreted as emission from the shocks on either side of the contact discontinuity. Here we report on the first time-dependent modelling of the non-thermal emission in eta Car. We find that emission from primary electrons is likely not responsible for the gamma-ray emission, but accelerated protons interacting with the dense wind material can explain the observations. In our model, efficient acceleration is required at both shocks, with the primary side acting as a hadron calorimeter, whilst on the companion side acceleration is limited by the flow time out of the system, resulting in changing acceleration conditions. The system therefore represents a unique laboratory for the exploration of hadronic particle acceleration in non-relativistic shocks.
Westerlund 1 (Wd 1) is the most massive stellar cluster in the Galaxy and associated with an extended region of TeV emission. Here we report the results of a search for GeV gamma-ray emission in this region. The analysis is based on ~4.5 years of Fer
mi-LAT data and reveals significantly extended emission which we model as a Gaussian, resulting in a best-fit sigma of sigma_S = (0.475 +/- 0.05) deg and an offset from Wd 1 of ~1 deg. A partial overlap of the GeV emission with the TeV signal as reported by H.E.S.S. is found. We investigate the spectral and morphological characteristics of the gamma-ray emission and discuss its origin in the context of two distinct scenarios. Acceleration of electrons in a Pulsar Wind Nebula provides a reasonably natural interpretation of the GeV emission, but leaves the TeV emission unexplained. A scenario in which protons are accelerated in or near Wd 1 in supernova explosion(s) and are diffusing away and interacting with molecular material, seems consistent with the observed GeV and TeV emission, but requires a very high energy input in protons, ~10^51 erg, and rather slow diffusion. Observations of Wd 1 with a future gamma-ray detector such as CTA provide a very promising route to fully resolve the origin of the TeV and GeV emission in Wd 1 and provide a deeper understanding of the high-energy (HE) astrophysics of massive stellar clusters.
Non-thermal hard X-ray and high-energy (HE; 1 MeV < E < 100 GeV) gamma-ray emission in the direction of Eta Carina has been recently detected using the INTEGRAL, AGILE and Fermi satellites. So far this emission has been interpreted in the framework o
f particle acceleration in the colliding wind region between the two massive stars. However, the existence of a very fast moving blast wave which originates in the historical 1843 Great Eruption provides an alternative particle acceleration site in this system. Here we explore an alternate scenario and find that inverse Compton emission from electrons accelerated in the blast wave can naturally explain both the flux and spectral shape of the measured hard X-ray and HE gamma-ray emission. This scenario is further supported by the lack of significant variability in the INTEGRAL and Fermi measured fluxes.
