Seminar AI for Environmental Sustainability

With the advent of satellite missions providing increasing volumes of publicly available satellite data (e.g. Copernicus program, 2014), the field of Earth Observation (EO) has embraced the Deep Learning revolution, developing modern and efficient models capable of supporting numerous decision-making applications. These solutions cover a wide and ever growing range of use-cases such as deforestation monitoring, natural disaster management, or even air quality and climate tracking, helping governments and organizations make science-based decisions when designing policies. The aim of this seminar is to learn about state-of-the-art research in AI for Environmental Sustainability, with an emphasis on Earth Observation-enabled solutions. We will also discuss the benefits, and technical or ethical limitations of the methodologies encountered.

The participants of the seminar will choose a topic of interest, based on a few proposals made by the teaching staff at the beginning of the class. Each participant will proceed with a review of the literature around their topic, and present their findings during the weekly sessions, in the form of a talk. Besides, a written report will be submitted by the participants mid-term, or at the end of the semester.

Have good knowledge of Data Science, ideally acquired through the courses by Prof. Lindauer and Prof. Rosenhahn. Additional knowledge in modern Computer Vision methodologies will be beneficial.