Neural Networks and Deep Learning
License: XG85CWJ4JWDN (July 2018)
Introduction course to deep learning and neural networks. The course covers logistic regression, shallow and deep (L-layered) neural networks, forward and back-propagation, cost computation and different activation functions.
Improving Deep Neural Networks
License: WE762YKYH2MJ (July 2018)
This course builds up on the previous one and focuses on using regularization techniques such as L2, dropout, data augmentation and early stopping as well as optimization methods like mini-batch gradient descent, gradient descent with momentum, Adam Optimization, learning-rate decay and batch normalization. This course also presents TensorFlow as a more efficient alternative to just Python and Numpy.
Structuring Machine Learning Projects
License: 6GKUMXNV7CAT (August 2018)
This course covers error analysis and how to prioritize the best directions to reduce it, the concepts of data mismatch and artificial data synthesis as well as applying transfer learning and multi-task learning for problems with small data sets or solving for multiple tasks at once.
Convolutional Neural Networks
License: 4B2CTVS5YDWH (February 2019)
Introduction to CNNs and residual networks. The main subjects covered are object recognition and localization using anchor boxes and the YOLO algorithm, solving facial verification and recognition problems and generating artistic renderings of images using neural style transfer. The course also introduces Keras as a high-level abstraction from Tensorflow for the programming assignments.
Sequence Models
License: 86HHBEKWFYMA (February 2019)
Introduction to sequential models and Recurrent Neural Networks (RNN) as well as GRU and LSTMs, applied mainly to NLP problems. The main points covered are text generation, audio transcription, text translation and trigger word detection.
