Factors Enabling the Breakthrough of Artificial Intelligence – Part 1

It is a common thing these days to hear things about Artificial Intelligence (AI). But what are the factors that have enabled AI to become a household name that it is today? These are namely Compute, Data, Tools, and Momentum. Let us take a detailed look at each of them.

Compute

The amount of computing required to do this stuff hasn’t existed until very recently. Between 1970 and 2000, the number of transistors on an integrated circuit board grew by about 50,000 and today, the number is close to 20,000,000,000. GPUs are the kings of compute for machine learning tasks, and that dedicated Tensor Processing Units are already being manufactured.  The exponential growth of the compute resources required to support machine isn’t just limited to computing power – it applies to storage, power efficiency, cost, and size of components. By reducing the cost of entry to the field, cloud compute resources have enabled machine learning research. Seriously powerful computing resources are available to the public, for rent, and at a low cost.

Data Availability

If we take a rough estimate, ninety percent of all data that exists today was created within the last two years. It was estimated that roughly 2.5 quintillion bytes of new data were being generated daily, looking at figures from a year ago. Training a supervised machine learning model requires accurately labelled data. The more samples you have, the better chances you have at creating a useful and accurate model. Smartphones and social media have contributed majorly to the pool of data enabling some popular consumer-focused AI services over the past decade.