Campus, PAB 102/103
The mysteries of the cosmic beginning, gravitational clustering, and cosmic acceleration persist. How can we distill relevant cosmological information from the next generation of data sets? Taking examples from the cosmic microwave background, large scale structure, and supernova cosmology, I will discuss inference strategies, artificial intelligence, and computational approaches that promise to extract more information from current and upcoming data sets. The philosophy is to allow maximum freedom to design realistic forward models, to be robust to systematic nuisances, to accurately combine multiple probes, move beyond simplistic likelihood assumptions, naturally allow quantitative model comparison, characterize tensions in the data, and maintain (near-)optimality whenever possible.