Deep Learning3 [DL] Lecture6:Deep learning for computer vision Objectives Explain the main uses of deep learning in computer vision Design and implement deep learning systems for computer vision Tutorial Encoder - decoder one-stage detector and two-stage detector computer vision 분야는 크게 4개. Image recognition, semantic segmentation, object detection, instance segmentation 있다. PART1: Semantic Segmentation semantic segmentation (semantic 이란 단어는 언어나 논리에 따른 관계를 지.. 2023. 5. 6. [DL] lecture4: Optimization and Regularization Techniques Objectives: Explain and implement different types of optimisation algorithms based on variations of stochastic gradient descent (SGD). Explain and implement different types of regularization strategies. Design training schemes for deep networks that make appropriate and effective use of optimization algorithms and regularization strategies. Tutorial PMSProp parameter estimator algo L2-Norm regul.. 2023. 5. 5. [DL] Lecture3: Deep Feedforward Networks-시험공부 Objectives Define the structure of a deep feedforward network Be able to choose and justify different activation functions for hidden layers and the output layer. Explain the backpropagation algorithm Tutorial deep forward network를 수학적 표현 ReLu 를 수학적으로 표현. Part1:Deep feedforward network definition Feedforward network : 피드백을 받지않고 이전 레이어에서 다음 레이어로 앞으로 전달만해주는 뉴럴네트워크 개념. 첫번째 레이어의 뉴론은 두번째 레이어의 뉴론들에 다 .. 2023. 5. 5. 이전 1 다음
