Deep Learning 2.0

#artificialintelligence 

In a video generation system using Action Prediction (Figure 3), if you drop a drop a ball, the AI should predict the ball will fall on the floor because the implicit knowledges "gates" the model to generate a future where the ball goes upward like a Helium balloon. But how to build such "gating mechanism" (seen in Figure 4) that uses cognitive knowledge of cause & effect? Is it possible to design a futuristic AI that can be trusted by a doctor for a AI diagnosis inspite of sample from a out of distribution? Will mere selective prediction suffix or a deeper understanding of the semantic context and understanding of human anatomy activate a medical decision. During medical diagnosis, how can an AI avoid a false prediction based on conceptual knowledge of the medical domain?

Duplicate Docs Excel Report

Title
None found

Similar Docs  Excel Report  more

TitleSimilaritySource
None found