Galaxy formation and evolution

Abstract:

HARMONI is the first light visible and near-infrared (NIR) integral field spectrograph for the Extremely Large Telescope(ELT). The HARMONI spectrograph will have four near-infrared cameras and two visible, both with seven lenses of various materials and diameters ranging from 286 to 152 mm. The lens mounts design has been optimized for each lens material to compensate for thermal stresses and maintain lens alignment at the operational temperature of 130 K. We discuss their design and mounting concept, as well as assembly and verification steps. We show initial results from two prototypes and outline improvements in the mounting procedures to reach tighter lens alignments. To conclude, we present a description of our future work to measure the decentering of the lenses when cooled down and settled.

Abstract:

HARMONI is the first light visible and near-IR integral field spectrograph for the ELT. It covers a large spectral range from 450nm to 2450nm with resolving powers from R (≡λ/Δλ) 3500 to 18000 and spatial sampling from 60mas to 4mas. It can operate in two Adaptive Optics modes - SCAO (including a High Contrast capability) and LTAO - or with NOAO. The project is preparing for Final Design Reviews. The instrument uses a field splitter and image slicer to divide the field into 4 sub-units, each providing an input slit to one of four nearly identical spectrographs. This proceeding presents the final opto-mechanical design and the AIV plan of the spectrograph units.

Abstract:

Computer simulations are invaluable tools for scientific discovery. However, accurate simulations are often slow to execute, which limits their applicability to extensive parameter exploration, large-scale data analysis, and uncertainty quantification. A promising route to accelerate simulations by building fast emulators with machine learning requires large training datasets, which can be prohibitively expensive to obtain with slow simulations. Here we present a method based on neural architecture search to build accurate emulators even with a limited number of training data. The method successfully accelerates simulations by up to 2 billion times in 10 scientific cases including astrophysics, climate science, biogeochemistry, high energy density physics, fusion energy, and seismology, using the same super-architecture, algorithm, and hyperparameters. Our approach also inherently provides emulator uncertainty estimation, adding further confidence in their use. We anticipate this work will accelerate research involving expensive simulations, allow more extensive parameters exploration, and enable new, previously unfeasible computational discovery.