DL for CV

Learningbuilding artificial systems that learn from data and experience Deep LearningHierarchical learning algorithms with many layers, very loosely i

input: image —> output : assign image to one of a fixed set of categories fine-grained categories no obvious way to hard-code the algorithm for recogn

RNN은 순차적인 데이터를 처리하기 위해 개발되었으며, 이는 feedforward network의 한계를 극복variable-length sequences 처리 가능 capturing temporal dependenciesrecurrent neural networks의

C : encoded sequence와 decoded sequence 사이의 정보를 전딜하는 과정 문제 : input sequence bottlenecked through fized-sized vector. What if T=1000? C (context vector)