Data operations concur greatly to the cost and delivery time for AI solutions.

Particularly the process of teaching the algorithm a certain task, the learning, requires significant human intervention.

A particular element of this intervention is the labeling of data. This process is extremely manual and constitute one of the biggest taxes to pay to develop Ai systems.

In recent years a new technique is emerging to alleviate this burden. Particularly it is the ability that the intelligence it self can modulate the amount of data labels that it requires during the learning process.

This new trend is described in this article:

https://aws.amazon.com/blogs/apn/the-curse-of-big-data-labeling-and-three-ways-to-solve-it/