
Sign up to save your podcasts
Or


Deep Learning - Loss and Optimization Part 2
This video explains hinge loss and its relation to support vector machines. It enables to embed optimization constraints into loss functions.
Further Reading:
Deep Learning - Loss and Optimization Part 3
This video discusses details on optimization and different options in gradient descent procedure such as momentum and ADAM.
References
[1] Christopher M. Bishop. Pattern Recognition and Machine Learning (Information Science and Statistics). Secaucus, NJ, USA: Springer-Verlag New York, Inc., 2006.
Further Reading:
Deep Learning - Activations, Convolutions, and Pooling Part 1
This video presents the biological background of activation functions and the classical choices that were used for neural networks.
Further Reading:
Deep Learning - Activations, Convolutions, and Pooling Part 2
This video presents a variety of modern activation functions including the quest for finding new ones.
Further Reading:
Deep Learning - Activations, Convolutions, and Pooling Part 3
This video presents convolutional layers including the concepts of strided and dilated convolutions and how to compute their gradient.
Video References:
Further Reading:
Deep Learning - Activations, Convolutions, and Pooling Part 4
This video presents max and average pooling, introduces the concept of fully convolutional networks, and hints on how this is used to build deep networks.
References:
Further Reading:
Deep Learning - Regularization Part 1
This video discusses the problem of over- and underfitting. In order to get a better understanding, we explore the bias-variance trade-off and look into the effects of training data size and number of parameters.
Further Reading:
From the publisher's feed
(multilayer) perceptron, backpropagation, fully connected neural networks
loss functions and optimization strategies
convolutional neural networks (CNNs)
activation functions
regularization strategies
common practices for training and evaluating neural networks
visualization of networks and results
common architectures, such as LeNet, Alexnet, VGG, GoogleNet
recurrent neural networks (RNN, TBPTT, LSTM, GRU)
deep reinforcement learning
unsupervised learning (autoencoder, RBM, DBM, VAE)
generative adversarial networks (GANs)
weakly supervised learning
applications of deep learning (segmentation, object detection, speech recognition, ...)
The accompanying exercises will provide a deeper understanding of the workings and architecture of neural networks.

0 Listeners

0 Listeners

2 Listeners

0 Listeners

1 Listeners

0 Listeners

6 Listeners

0 Listeners

0 Listeners