My name is Aleksandr. I'm currently learning Computer Vision, especially GAN's.
Here is a few links you can take a look at:
Peak load reduction using thermal energy storage in a HVAC system of a building
| Project | Comment |
|---|---|
| Deep Convolutional Generative Adversarial Network | Created Generative Adversarial Network, which was trained at the MNIST-Fashion dataset. Generative NN training is shown in the GIF-fle with outputs after each epoch. |
| Neural Style Transfer using OpenCV | Attempt to implement the artistic style of a painting on a video using the OpenCV library. Also applied a style transfer to streaming video from the camera. |
| Web-page with printed Neural Network layers outputs | Created flask server, where at each request, a random instance is selected from the MNIST dataset and passed through the neural network. Returns the output values of each layer. Web-interface to visualise layers outputs was created using streamlit package. |
| Building of Docker container for flask app | Created flask server with image, where regression was plotted in temperature data. Application was builded to container using Dockerfile. |
| Outlier detection in financial data | Using of visualisation and clusterisation methods to find anomalies in unlabeled data. Tried Dash visualisation for .csv data-file. Packed in Docker. |
| McKinsey ProHack competition | Prediction of the development index of "galaxies" using regression, solving the problem of optimal resource allocation between them. Initial data distribution is asymptotic, also it has NaNs. I wrote the pipeline myself. Top-40% solution. Used one-hot encoding, tried classical regression in combination with one-layer NN (pytorch). Optimization problem solved by own algorithm, checked using CVXPY. |
| Title | Author |
|---|---|
| Introduction to Machine Learning | HSE/YandexDataAnalysisSchool |
| Deep Learning Specialisation | deeplearning.ai |
| SQL for Data Science | University of California, Davis |
| Version Control with Git | Atlassian |