This course teaches the concepts and theory behind convolution for one-dimensional data, specifically for time series. The learning outcomes include understanding 1D convolution for neural networks, implementing convolution in Python from scratch, creating convolution blocks, initializing blocks, and building a 1D convolutional neural network. The course covers topics such as sliding dot product, convolution equations, backpropagation, padding, stride, weight gradients, input gradients, and multi-channel convolutions. The teaching method involves a sequence of videos, practical implementations, and building and evaluating a 1D convolutional neural network. The intended audience for this course includes individuals interested in deep learning, neural networks, and specifically those working with one-dimensional data like time series.
Overview
Syllabus
1D convolution for neural networks, part 1: Sliding dot product.
1D convolution for neural networks, part 2: Convolution copies the kernel.
1D convolution for neural networks, part 3: Sliding dot product equations longhand.
1D convolution for neural networks, part 4: Convolution equation.
1D convolution for neural networks, part 5: Backpropagation.
1D convolution for neural networks, part 6: Input gradient.
1D convolution for neural networks, part 7: Weight gradient.
1D convolution for neural networks, part 8: Padding.
1D convolution for neural networks, part 9: Stride.
Implement 1D convolution, part 1: Convolution in Python from scratch.
Implement 1D convolution, part 2: Comparison with NumPy convolution().
Implement 1D convolution, part 3: Create the convolution block.
Implement 1D convolution, part 4: Initialize the convolution block.
Implement 1D convolution, part 5: Forward and backward pass.
Implement 1D convolution, part 6: Multi-channel, multi-kernel convolutions.
Implement 1D convolution, part 7: Weight gradient and input gradient.
Build a 1D convolutional neural network, part 1: Create a test data set.
Build a 1D convolutional neural network , part 2: Collect the Cottonwood blocks.
Build a 1D convolutional neural network , part 3: Connect the blocks into a network structure.
Build a 1D convolutional neural network, part 4: Training, evaluation, reporting.
Build a 1D convolutional neural network, part 5: One Hot, Flatten, and Logging blocks.
Build a 1D convolutional neural network, part 6: Text summary and loss history.
Build a 1D convolutional neural network, part 7: Evaluate the model.
Taught by
Brandon Rohrer
