Siamese Networks
Eric delves into the concept of siamese networks, explaining how they share weights to create a single vector representation of input images. Tim discusses the use of contrastive loss in neural network design, emphasizing the shift from reconstruction to distinguishing inputs based on distance metrics.In this clip
From this podcast
Machine Learning Street Talk (MLST)
One Shot and Metric Learning - Quadruplet Loss (Machine Learning Dojo)
Related Questions