CRCV REU 2019 Aaron Honculada

Theory Deep Learning Artificial Neural Network (ANN) Backpropagation Gradient Descent Loss Function Activation Function Hyper Parameters Useful for classification

Theory Convolutional Neural Network (CNN) Convolution 3D Convolution Residual Network Inception Model [1] Learning Spatiotemporal Features with 3D Convolutional Networks

Theory Recurrent Neural Network (LSTM) Auto-Encoder Sequential Data Auto-Encoder Generative Adversarial Network (GAN) MAC cell [1] COMPOSITIONAL ATTENTION NETWORKS FOR MACHINE REASONING

Models ResNet I3D [1] Deep Residual Learning for Image Recognition [2] Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning

Thanks to Robert, Aisha, and Kartik