Neural networks

September 21, 2018

Inspired by the working of the brain, neural networks are computer programs made up of simple, highly connected elements that learn the underlying relationships within data through considering examples.

Mia Learning uses a type called deep convolutional neural networks. These networks use many layers of filters that identify increasingly abstract features within data, culminating in identification of the feature of interest, such as a particular named entity within a text. The network learns the optimal values for the filters and the weights of connections between them through training.

Share this article
Share On Facebook
Share On Twitter
Share On Linkedin
Share On Pinterest