Neural Network Training (recognizing objects)

Brief description

The user can explore how a neural network interpret objects by taking an object and bringing it in front of a camera. Large screens visualize how output values of one layer become input values for the next one and how the neural network convolutes to a final output value. Gradually the input image becomes abstract and unrecognizable for the human eye. This value determines with what confidence level the object is recognized as an object by the trained machine learning model. On the last screen the confidence level is displayed as the system classifies the object based on the dataset and taxonomy it has been trained with. 

Work that the situation appears in

Who does what?
This entity
This entity
This entity
This entity
is/are
Aesthetic characteristics
Machine P.O.V
Machine P.O.V.
Notes
I chose Machine P.O.V because the installation tries to demonstrate how machines see.

Authored by

UUID
0347b4bb-c1e1-460a-a814-b3109ee07bc6