
An Astronomer Does Machine Learning
In the past decade, deep learning has revolutionized machine perception and gone from being a little-known niche field to vastly outperforming other algorithms at tasks as diverse as image classification, text-to-speech production, and natural language processing. The tremendous success of deep learning has led to billions of dollars of investment in the development of these models in industry. Nevertheless, foundational open questions remain in the field and I argue these are inherently scientific in nature. I will provide a brief introduction to machine learning for astronomers and discuss what I see are the outstanding open questions, along with a particular application of these techniques to the problem of denoising in hearing aids.