Scribd
Upload
102 views19 pages

Neural - Data - Science - 0 Introduction

The document provides an introduction to the Neural Data Science course taught by Prof. Dr. Philipp Berens, including topics covered such as models of neural circuits and computations, single cell transcriptomics, and AI-based medical diagnostics. It outlines the goals of learning common problems in neural data science, analyzing different data types, and implementing algorithms to solve these problems. The course involves programming exercises, a report, lectures, and tutorials to provide students with working knowledge of important techniques and literature in the field.

Uploaded by

fdasffd
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
102 views19 pages

Neural - Data - Science - 0 Introduction

The document provides an introduction to the Neural Data Science course taught by Prof. Dr. Philipp Berens, including topics covered such as models of neural circuits and computations, single cell transcriptomics, and AI-based medical diagnostics. It outlines the goals of learning common problems in neural data science, analyzing different data types, and implementing algorithms to solve these problems. The course involves programming exercises, a report, lectures, and tutorials to provide students with working knowledge of important techniques and literature in the field.

Uploaded by

fdasffd
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Neural Data Science

Introduction
Prof. Dr. Philipp Berens
Institute for Ophthalmic Research, Tübingen
[Link]

2 | Philipp Berens
Topics & interests of the lab

• Models for understanding circuits


& computations in the retina

• Single cell transcriptomics in


neuroscience

• AI-based medical diagnostics in


ophthalmology

3 | Philipp Berens
4 | Philipp Berens
Neural data
Ecker et al. 2010

Cadwell et al. 2016

Baden et al. 2016

Wikipedia Jiang et al. 2015


Multimodal data

Scala, Kobak et al. Nature 2020


Multimodal data

[Link]
Why is there an extra course?
• Machine Learning
• Signal Processing
• Neural Coding
• Neural Dynamics
An experimentalist’s ideal

Scientific
Experiment Data
question

Write a
paper
An alternative view

Scientific
Experiment Data
question

Data
analysis

Write a
paper
How Does Data Analysis Work?

Scientific Data
question Data
analysis

Machine
Learning

Signal
Neural
Processing
Coding

TOOLBOX
Learning targets
• Learn about common problems in neural data science
• Analyse different data types
• Implement and apply basic algorithms to solve them
• Learn about advanced solutions
• Learn how to analyse complex real world datasets

• Practise a data science philosophy


- Look at raw data
- Look at examples
- Test algorithms on artificial data
- Be critical with your results
- Let the problem determine the technique

12 | Philipp Berens
What knowledge is helpful for the course?
• Python programming

• Basic machine learning


knowledge
- Clustering
- Regression
- Classification

• Some neuroscience context

14 | Philipp Berens
How does this course work? (1/2)
• Lecture (online on the weekend, short)
- Problem statement & possible solutions
- Algorithmic details and application issues
- Goal: working knowledge of important problems &
techniques, literature overview

• Tutorial (Thursday, 2.00pm-3.30pm)


- Held by me or a tutor
- Tutors: Sarah Strauss & Ziwei Huang
- Goals: get answers to questions, discuss solution to
exercises
- Be prepared!
How does this course work? (2/2)
• Programming exercises (75% of grade)
- Work on real-world problems and do analysis
- Implement algorithms discussed in lecture as model solution

• Report (25% of grade)


- Detailed report about one exercise, choose on Ilias
- Problem statement, solutions, well-prepared figures, results,
brief discussion of literature
- Template will be provided

• No exam
• 4 ECTS version: Only programming exercises, no report!
Programming Exercises
You can work in groups of two students
Due Mondays, 8.15 AM

1. Jupyter notebooks and data provided on ILIAS


2. Upload pdf and Jupyter notebook to ILIAS
3. In tutorial:
- Discussion about solution and implementation
- You should be able to explain the figures and the code
- Have your notebooks ready

The programming exercise is a serious time commitment!


Policies
• No late submissions.
• Use any available material, including Python libraries (unless we
specifically ask you not to do this).
• Talk to others about the problems, but
- credit any source in the code
- write your own code.
• Work as a team, provided
- you disclose who did what
- every group member can explain each part.

• If you have feedback how to improve the course, talk to us.


Write good code
• Keep it simple!
• Write as little and clean code as possible.
• Comment your code, but don’t overdo it.
• Don’t worry about efficiency. Readability is more important.
• Keep to the interface!

You might also like