Green AI Seminar

Content

This seminar explores the potential of AI in combating global climate change by discussing its role in monitoring climate crises, conserving nature, and reducing greenhouse gas emissions in various sectors. The seminar will also consider the sustainability of AI itself, as advances in machine learning come with a significant increase in carbon emissions. Participants will learn about essential aspects of improving the sustainability of machine learning algorithms and gain a different perspective on machine learning’s role in addressing climate change.

 

The seminar provides:

    • Introduction to “Green AI” versus “Red AI”
    • Guests talks on related research topics
    • Group discussions on future prospects of AI, specifically machine learning
    • Best practices for literature review and scientific presentations
    • Literature review on Green AI in certain areas in groups
    • Scientific talk of each student on one specific topic

 

Contact: Dr.-Ing. Eva Dorschky and René Raab

Time and place: Wednesday 10:15 – 11:45, MaD seminar room (ground floor, Carl-Thiersch-Str. 2b, 91052 Erlangen)

 

Guest Talks

Solar energy forecasting from cloud cover observations

This talk will present recent advances in solar forecasting with computer vision, with a focus on how machine learning can be used to anticipate the future solar power variability from cloud cover observations. Accurately predicting future changes in solar energy production is key to facilitate the large-scale integration of this low-carbon energy source into power systems.

 

Guest speaker: Quentin Paletta (European Space Agency)

Date: 22.11.2023

Zoom only: Zoom link, Meeting ID: 698 5376 3200, Passcode: 131943

 

Forest fire detection using drones

This talk will present research on the use of convolutional neural networks (CNN) to automatically search for forest- and wildfires in drone-based applications. It will be shown how the combination of AI with autonomous long distance drones can make life safer and easier for fire brigades and dampen the effects of wildfires on climate change.

Guest Speakers: Isabella Hufnagl and Simon Gehringer (Evolonic)

Date: 13.12.2023

 

 

Neuromorphic Computing: The state and potential of Spiking Neural Networks

This talk will present an introduction to biology-inspired Spiking Neural Networks (SNNs), their current state, and their potential as an energy-efficient alternative to conventional artificial neural networks (ANNs). In addition, the necessity, benefit, and challenges of developing neuromorphic hardware suitable for SNNs will be discussed.

 

Guest speaker: Ferdinand Pscheidl (Fraunhofer EMFT)

Date:  07.02.2024

Zoom only: Zoom link, Meeting ID: 683 7186 9072, Passcode: 568984

 

 

Schedule

Seminar Dates
Introduction and group work 18.10.2023
Literature review and scientific writing 25.10.2023
Scientific presentations 08.11.2023
Sustainability of AI 15.11.2023
Invited talk by Quentin Paletta on solar energy forecasting 22.11.2023
Q & A 29.11.2023
Interim report of groups 06.12.2023
Invited talk by Evolonic on forest fire detection using drones 13.12.2023
Student talks 20.12.2023
Student talks 10.01.2024
Student talks 17.01.2024
Student talks 24.01.2024
Student talks 31.01.2024
Invited talk by Ferdinand Pscheidl on spiking neural networks 07.02.2024