Paul Miller

SARS-CoV-2 Plots

Python, pandas, matplotlib, Tableau

Project to draw plots using the Swiss open government datasets collected on the OpenZH Github repository.

Covid_19
Map of Weekly Confirmed Cases
Map of Weekly Deaths
Other visualizations on Tableau Public

SARS-CoV-2 Vaccination Plots

Python, pandas, matplotlib

Project to draw plots using the FOPH vaccination data.

FOPH_covid_19

Blood cells

Python, pytorch, numpy, matplotlib

Classify blood cell images using a pretrained ResNet50 model as a feature extractor.

Blood_cells

Fruits360

Python, pytorch, numpy, matplotlib

Classify fruit images using a pretrained VGG16 model as a feature extractor.

Fruits360

Predict Future Sales

Python, pytorch, pandas, numpy, Tableau

This is my approach to the "Predict Future Sales" competition from Kaggle. The goal of the competition is to predict sales for different items for the next month based on past data.

PredictFutureSales

Face recognition

Python, opencv, dlib, face_recognition

Face recognition implementation which uses the face_recognition library to recognize people in a video based on photos of them.

Face_rec

Document scanner

Python, opencv, pytesseract

Scan document and run optical character recognition to get raw text.

document-scanner

MNIST GAN

Python, pytorch

A Deep Convolutional Generative Adversarial Network which generates MNIST images to train a discriminator.

DC_GAN

Aggregation of Dropout Subnetworks

Python, pytorch, LaTeX, numpy, pandas

My Bachelor's project consisted on working on the Dropout method used to reduce overfitting in artifical neural networks. Specifically, I tried to find a better way to aggregate the subnetworks that the Dropout technique creates by applying it at test time and comparing the accuracy between different aggregation techniques.

bachelor
DD_dropout

Snake

Python, pytorch

Create a snake to train itself through Deep Reinforcement Learning.

Snake

Titanic

Python, pytorch, pandas, Tableau

This is my approach to the "Titanic: Machine Learning from Disaster" competition from Kaggle. The goal of the competition is to predict which passengers survived the shipwreck.

Titanic

WikiScrape

Python, pandas

A small module to scrape Wikipedia tables and output Pandas dataframes.

wikiscrape