August 2019
Intermediate to advanced
342 pages
9h 35m
English
The centrality of data in the correct functioning of data-driven solutions is attested by the principle known as garbage in, garbage out. Protecting the integrity of data is equally important in the AI, since—unlike what is commonly thought—algorithms are not objective, but can formulate radically different predictions based on the data supplied to them in the training phase.
It is therefore possible to condition the results of predictive models simply by altering the data on which the algorithms work.
This aspect is particularly delicate, as it transversally raises the need to guarantee the security and integrity of the data, together with the explicability and repeatability of the results obtained by ...
Read now
Unlock full access