We used data from the near-infrared Visible and Infrared Telescope for Astronomy (VISTA) survey of the Magellanic Cloud system (VMC) to measure proper motions (PMs) of stars within the Small Magellanic Cloud (SMC). The data analysed in this study comprise 26 VMC tiles, covering a total contiguous area on the sky of similar to 40deg(2). Using multi-epoch observations in the K-s band over time baselines between 13 and 38 months, we calculated absolute PMs with respect to similar to 130000 background galaxies. We selected a sample of similar to 2160000 likely SMC member stars to model the centre-of-mass motion of the galaxy. The results found for three different choices of the SMC centre are in good agreement with recent space-based measurements. Using the systemic motion of the SMC, we constructed spatially resolved residual PM maps and analysed for the first time the internal kinematics of the intermediate-age/old and young stellar populations separately. We found outward motions that point either towards a stretching of the galaxy or stripping of its outer regions. Stellar motions towards the North might be related to the 'Counter Bridge' behind the SMC. The young populations show larger PMs in the region of the SMC Wing, towards the young Magellanic Bridge. In the older populations, we further detected a coordinated motion of stars away from the SMC in the direction of the Old Bridge as well as a stream towards the SMC.

Numerous studies have demonstrated the ability of Convolutional Neural Networks (CNNs) to classify large numbers of galaxies in a manner that mimics the expertise of astronomers. Such classifications are not always physically motivated, however, such as categorizing galaxies by their morphological types. In this work, we consider the use of CNNs to classify simulated S0 galaxies based on fundamental physical properties. In particular, we undertake two investigations: (1) the classification of simulated S0 galaxies into three distinct evolutionary paths (isolated, tidal interaction in a group halo, and spiral spiral merger) and (2) the prediction of the mass ratio for the S0s formed via mergers. To train the CNNs, we first run several hundred N-body simulations to model the formation of S0s under idealized conditions, and then we build our training data sets by creating images of stellar density and two-dimensional kinematic maps for each simulated S0. Our trained networks have remarkable accuracies exceeding 99 per cent when classifying the S0 formation pathway. For the case of predicting merger mass ratios, the mean predictions are consistent with the true values to within roughly one standard deviation across the full range of our data. Our work demonstrates the potential of CNNs to classify galaxies by the fundamental physical properties that drive their evolution.

In the rest frame of the Local Group (LG), the total momentum of the Milky Way (MW) and Andromeda (M31) should balance to zero. We use this fact to constrain new solutions for the solar motion with respect to the LG centre of mass, the total mass of the LG, and the individual masses of M31 and the MW. Using the set of remote LG galaxies at > 350 kpc from the MW and M31, we find that the solar motion has amplitude V-circle dot = 299 +/- 15 km s(-1) in a direction pointing towards galactic longitude l(circle dot) = 98.degrees 4 +/- 3.degrees 6 and galactic latitude b(circle dot) = -5.degrees 9 +/- 3.degrees 0. The velocities of M31 and the MW in this rest frame give a direct measurement of their mass ratio, for which we find log(10)(M-M31/M-MW) = 0.36 +/- 0.29. We combine these measurements with the virial theorem to estimate the total mass within the LG as M-LG = (2.5 +/- 0.4) x 10(12) M-circle dot. Our value for M-LG is consistent with the sum of literature values for M-MW and M-M31. This suggests that the mass of the LG is almost entirely located within the two largest galaxies rather than being dispersed on larger scales or in a background medium. The outskirts of the LG are seemingly rather empty. Combining our measurement for M-LG and the mass ratio, we estimate the individual masses of the MW and M31 to be M-MW =(0.8 +/- 0.5) x 10(12) M-circle dot and M-M31 =(1.7 +/- 0.3) x 10(12) M-circle dot, respectively. Our analysis favours M31 being more massive than the MW by a factor of similar to 2.3, and the uncertainties allow only a small probability (9.8 per cent) that the MW is more massive. This is consistent with other properties such as the maximum rotational velocities, total stellar content, and numbers of globular clusters and dwarf satellites, which all suggest that M-M31/M-MW > 1.