| Learning about various neural network (NN) architectures (k4): Students will learn about the concepts and implementation details of feedforward NNs, convolutional NNs (CNN), recurrent NNs (RNN), and autoencoders
(AE).
Building basic feedforward neural networks from scratch (k5): Students can implement a basic form of a
feedforward neural network from the ground up and train these models on complex datasets.
Implementing and training advanced NNs (k5): Students are able to design and train various modern NN types such as CNNs, RNNs, and AEs on tasks involving real-world data using state-of-the-art deep learning code libraries.
Evaluating and Optimizing NNs (k5): Students can evaluate model performance using appropriate metrics
and optimize hyperparameters such as learning rates and batch sizes. Students can also use regularization
techniques to minimize overfitting.
|
Students have practical knowledge of implementing advanced neural network models, including CNNs,
RNNs, and AEs, using modern Python libraries. They have learned to apply these models to real-world tasks,
experiment and solve challenges in training and optimization, and gain insights into the practical applications of deep learning.
|
