Champagne Coding: Image Recognition
Details
Our human brain makes vision seem like an easy task, but this is a complex problem for computers. In the last decade, researchers have made great progress in computer vision by training deep learning models on millions of images.
In this session, Ning Zhou will be joining us with her colleagues Audun and Bjørn to explain why deep learning works for images and show you how to use transfer learning to train your own model with just a couple hundred images. Ning, Audun and Bjørn had a great time working together on marketplace images from their time at Schibsted and are looking forward to discussing all things image with you!
In the hands-on session, we will use Keras, a neural network library in Python, to train an image recognition model for flowers. You will then be able to test your model to identify flower types - is it a dandelion or a daisy? or maybe it's a tulip? We will also discuss the limitations of the model and potential solutions.
------
Prerequisites: A laptop with Anaconda installed. Basic knowledge in Python (e.g. know how to write a for loop) and machine learning (e.g. understand what training means) are strongly recommended but not required.
------
The event will be held at Wergelandsveien 7 in the Arundo offices. From 5-6, we will serve finger food and drinks. From 6 - 8, we will run the workshop (with champagne!). If you have a food allergy, please send a message and we will try to accommodate!
------
Women in Data Science is a group designed to inspire and educate data scientists worldwide, regardless of gender, and support women in the field.
