Data Science Retreat Demo Day #22

Data Science Retreat presents 9 Machine Learning prototypes and projects by Batch 22/23 participants. The event is free to attend but please let us know you are coming. We are excited to organise the first in-person demo day for the year.

17:00 Drinks and Networking

17:30 Welcome and Introduction

17:40 Project Presentation
“Deep Learning for Crypto market Prediction Using NLP and Technical Indicators”
Project by - Philip Paprotny, Miguel Gonzalez, Matias Lueska

The project uses Deep Learning models (LSTM) for daily directional movements prediction of crypto prices using financial news, social media and technical indicators as input.

18:00
“Monitoring water levels with satellite imagery”
Project by - Maximiliano Costa Beber and Karl Radtke

The goal is to use deep learning image segmentation algorithms and satellite imagery to asses changes to freshwater over time.

18:15
“Deep Learning for Sign Language translation”
Project by - Tomasz Iżycki

That project tries to make a bridge between deaf and hearing people without their knowledge of sign language. By using conv-nets, openpose and LSTM we are able to transform a video into a transcription.

18:30
“Deep food”
Project by - Gleb Moiseenko, Iskriyana Vasileva, Michael Drews, Nima Siboni

Make a photo of your fridge, recognize all ingredients and generate matching recipes with deep learning. Our model combines transfer learning and object recognition networks, trained on artificially and automatically generated data.

18:50 Break

19:00
“Gentri-Watch”
Project by - Katharina Thumser

Using clustering and time series analysis to predict the risk of gentrification for Berlin neighborhoods.

19:15
“Plum Music”
Project by - Christopher Dick

This project is working to generate music and transfer music styles, leveraging deep learning in creativity by using Google's Tensorflow Magenta. With state-of-the-art Machine Learning approaches (Variational Auto Encoders, Wavenet like Deep Nets, ...) the possibilities of creating and altering the creative art form of music is explored.

19:30
“Deeplexia”
Project by - Elena Stein Scholtis and Jerome D'Harveng
The project is a web app to convert a text to emojis as a reading aid for dyslexic children and to help people train for speed reading.

19:45
“Comparison of two segmentation methods for lakes datasets.”
Project by - Olga Dimitrova

20:00
Employees Attrition-Aton
Project by - Nadav Smith

Employee attrition is a pain that every business has. An employee is leaving in a very inopportune moment, or a key employee is choosing to move on can prove costly. Our model aims to give management and HR a tool to mitigate that risk. It assesses the likelihood that any given employee will leave. It also aims to pinpoint the reasons why that employee might churn, giving the management and HR insight into how to handle the situation ahead of time.