Google working to fix AI bias problems

Google has been using TCAV (testing with concept activation vectors) technology to help explain how an AI comes to a decision using higher-level human-friendly terms.


The use of TCAV aims to avoid perpetuating biases in society and data, as well as to inform experts when evaluating models.


Read here.



#Google #TCAV #Bias #Data #AI

©2019 by AITHICS