Join us for a GOTO Night with two great talks on Machine Learning with speakers Stefan Veistrup and Alexander Krog!
Machine Learning (ML) has been around for decades and really taken off within the last few years due to the advances within Deep Learning. A lot of people have heard about these terms and understand the possibilities with these technologies: “You use a lot of data to train a machine to do amazing stuff”.
This presentation won’t go too much in detail about the technicalities of ML, but rather give an overview of ML and some of the common considerations when starting a ML project. “Should I use IBM Watson? Should I try this ready-to-use model I found on GitHub? Should I train my own model?”. It will also cover some common problems that ML engineers face through easy to understand examples.
Stefan has a background as a mechatronics engineer, but has changed his focus to native mobile app development and machine learning. Currently working as an iOS and ML Engineer @ Trifork Aarhus.
Generating an appropriate feature representation is often one of the most important and challenging part of applying machine learning.
It doesn't matter how much data you have or which algorithm you use, if the relevant information in your data can’t be extracted it’s not going to work. We therefore need techniques for finding and representing relevant information in large data sets. Traditionally this is done using the manually feature engineering approach, which is based on domain knowledge and intuition.
Feature learning is techniques for automatically generating feature representations from data structures, avoiding the need for the time consuming feature engineering process.
This presentation will cover what feature learning is, and how it can be used to pre-process data for machine learning applications.
Alexander is a Phd student at AAU studying Machine and Feature Learning