Georgia Runoff Model Evaluation
Now that the election has ended its worth evaluating how well the model did. We are interested in both the final predictions after all polls were collected and how the model changed as new data came in.
Hierarchical Partial Pooling for Georgia Runoff Elections
Updated: Jan 5, 2021 (last update before results?)
Warnock: 69.3% - Loeffler: 30.4%
Ossoff: 58.6% - Perdue: 41.4%
This is a good...
Generating Death Metal Band names with Recurrent Neural Networks
For anyone familiar with Death Metal you know it’s a genre known for it’s extremes. The music, the names, the subject matter - it’s all extreme! I thought it would be interesting to see what a recurrent neural network would...
Reproducing CIFAR-10 results from deep and wide preactivation residual networks
In 2015, Deep Residual Networks [1] were introduced as the winning solutions to ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation, and they made it possible to train extremely deep neural networks of up to 1000 or more...
Measuring cardiac ejection fraction with deep convolution neural networks
I recently competed in the Second Annual Data Science Bowl which challenged participants to create a method to automatically measure end-systolic and end-diastolic volumes from cardiac MRIs. My goal was to see how far I could get...
Using deep learning to read street signs
Show me the code!
The First Steps with Julia competition on Kaggle uses a subset of the Chars74k dataset which consist of a series of characters cropped from natural images. Although the Kaggle competition was set up...
Neural network with numpy
Neural networks are a wonderful machine learning algorithm. They are called neural networks because they are loosely based on how the brain’s neurons work, which can make them seem intimidating. However, they are essentially a group of linear models. There...