Skip to content

WUT-ML/appliedmldays_2018

Repository files navigation

Discovering Brain Structure with Machine Learning

Introduction

This repository contains all the information regarding workshop: Discovering Brain Structure with Machine Learning that was delivered during the Applied Machine Learning Days organized at EPFL, Lausanne, Switzerland.

Goal

The goal of this 3 hours workshop is to faliliarize participants with classic machine learning techniques applied to the brain imaging data. Since, we do not assume neuroscience experience, all necessary concepts will be explained in plain english. Similarly, we do not assume any serious expertise in machine learning field. Our goal is to show ideas behind algorithms as well as present how such algotirhm can be used in the specific neuroscinece context.

Workshop schedule

note that schedule is the same for both Saturday and Sunday

Workshop is designed for 3 hours. It is self-contained in a sense, that we will cover all necessary concepts in machine learning and neuroscience. Thanks to this participants without prior experience in these fields may follow the contents.

time contents
9:00 - 9:15 Welcome remarks and introduction to the workshop
9:15 - 9:55 Introduction to neuroscience and machine learning
9:55 - 10:10 15 minutes break (good time to setup an environment for the technical part)
10:10 - 11:45 Hands-on part: analyse zebrafish imaging data (notebook)
11:45 - 12:00 Summary and closing remarks

Technical setup

There are two main ways to get started. Please refert to the Setup page on our Wiki.

Interactive notebooks

Hands-on part of this workshop will make use of the Jupyter notebooks. Notebook itself is an interactive execution environment that lets users evaluate single lines of code and observe an execution's output immediately. It is well-suited for exploratory and iterative work.

Data

Zebrafish imaging data is available for download here: dataset.

It was recorded in the Philipp Keller lab and is related to the publication: Whole-brain functional imaging at cellular resolution using light-sheet microscopy.