One of the principal bottlenecks to atmosphere characterisation in the era of all-sky surveys is the availability of fast, autonomous and robust atmospheric retrieval methods. We present a new approach using unsupervised machine learning to generate informed priors for retrieval of exoplanetary atmosphere parameters from transmission spectra. We use principal component analysis (PCA) to efficiently compress the information content of a library of transmission spectra forward models generated using the PLATON package. We then apply a $k$-means clustering algorithm in PCA space to segregate the library into discrete classes. We show that our classifier is almost always able to instantaneously place a previously unseen spectrum into the correct class, for low-to-moderate spectral resolutions, $R$, in the range $R~=~30-300$ and noise levels up to $10$~per~cent of the peak-to-trough spectrum amplitude. The distribution of physical parameters for all members of the class therefore provides an informed prior for standard retrieval methods such as nested sampling. We benchmark our informed-prior approach against a standard uniform-prior nested sampler, finding that our approach is up to a factor two faster, with negligible reduction in accuracy. We demonstrate the application of this method to existing and near-future observatories, and show that it is suitable for real-world application. Our general approach is not specific to transmission spectroscopy and should be more widely applicable to cases that involve repetitive fitting of trusted high-dimensional models to large data catalogues, including beyond exoplanetary science.

The large number of new planets expected from wide-area transit surveys means that follow-up transmission spectroscopy studies of their atmospheres will be limited by the availability of telescope assets. We argue that telescopes covering a broad range of apertures will be required, with even 1m-class instruments providing a potentially important contribution. Survey strategies that employ automated target selection will enable robust population studies. As part of such a strategy, we propose a decision metric to pair the best target to the most suitable telescope, and demonstrate its effectiveness even when only primary transit observables are available. Transmission spectroscopy target selection need not therefore be impeded by the bottle-neck of requiring prior follow-up observations to determine the planet mass. The decision metric can be easily deployed within a distributed heterogeneous network of telescopes equipped to undertake either broadband photometry or spectroscopy. We show how the metric can be used either to optimise the observing strategy for a given telescope (e.g. choice of filter) or to enable the selection of the best telescope to optimise the overall sample size. Our decision metric can also provide the basis for a selection function to help evaluate the statistical completeness of follow-up transmission spectroscopy datasets. Finally, we validate our metric by comparing its ranked set of targets against lists of planets that have had their atmospheres successfully probed, and against some existing prioritised exoplanet lists.

Recurrent novae (RNe) are novae with multiple recorded outbursts powered by a thermonuclear runaway. The outburst occurs on the surface of the white dwarf which accompanies with a late type main-sequence or giant secondary star transferring material onto the white dwarf primary star. They resemble classical novae (CNe) outbursts but only RNe has more than one recorded outbursts. RNe play an important role as one of the suspected progenitor systems of Type Ia supernovae (SNe) which are used as primary distance indicators in cosmology. Thus, it is important to investigate the outburst type of CNe and RNe and finally ascertain the population of objects that might ultimately be candidates for Type Ia SNe explosions. The proposal that RNe occupy a region separated from CNe in an outburst amplitude versus speed class diagram was adopted. Since the low amplitude results from the existence of an evolved secondary and/or high mass transfer rate in the quiescent system, RNe candidates should accordingly have low amplitude. We selected 3 preliminary targets including T Pyx, BT Mon and V574 Pup. Their amplitudes are not that low but the lowest amplitude that can be observed with Thai National Telescope (TNT). We obtained their magnitudes at quiescence using ULTRASPEC camera on the 2.4-m TNT. The positions of three targets on optical and near-infrared color-magnitude diagrams suggest that all three should have mainsequence secondary stars. This is true for T Pyx, whose secondary star has been confirmed its spectroscopy to be a main-sequence star, but not yet confirmed for BT Mon and V574 Pup.

SDSS J1021+1744 is a detached, eclipsing white dwarf / M dwarf binary discovered in the Sloan Digital Sky Survey. Outside the primary eclipse, the light curves of such systems are usually smooth and characterised by low-level variations caused by tidal distortion and heating of the M star component. Early data on SDSS J1021+1744 obtained in June 2012 was unusual in showing a dip in flux of uncertain origin shortly after the white dwarf's eclipse. Here we present high-time resolution, multi-wavelength observations of 35 more eclipses over 1.3 years, showing that the dip has a lifetime extending over many orbits. Moreover the "dip" is in fact a series of dips that vary in depth, number and position, although they are always placed in the phase interval 1.06 to 1.26 after the white dwarf's eclipse, near the L5 point in this system. Since SDSS J1021+1744 is a detached binary, it follows that the dips are caused by the transit of the white dwarf by material around the Lagrangian L5 point. A possible interpretation is that they are the signatures of prominences, a phenomenon already known from H-alpha observations of rapidly rotating single stars as well as binaries. What makes SDSS J1021+1744 peculiar is that the material is dense enough to block continuum light. The dips appear to have finally faded out around 2015 May after the first detection by Parsons et al. in 2012, suggesting a lifetime of years.