Revisiting Machine Learning

Imagine being transported back to the late 1990s in the United States, where the authorities discover your expertise in Machine Learning (ML). They reach out to seek your assistance in the automation of a time-consuming task: reading pin code on letters. Supposedly there are 500 such employees in various post offices across the country, and each employee was being paid a sum of $2000 per month to perform this task. This accumulates to a monthly expenditure of $1000000, resulting in an annual cost of $12 million, or a staggering $60 million nationwide, over the next five years.