Applications of Generative Adversarial Networks in Environmental Data Science

Add to your list(s) Download to your calendar using vCal

• Herbie Bradley | University of Cambridge
• Tuesday 02 February 2021, 11:00-12:30
• https://zoom.us/j/6708259482?pwd=Qk03U3hxZWNJZUZpT2pVZnFtU2RRUT09.

If you have a question about this talk, please contact Tudor Suciu.

Generative adversarial networks (GANs) are a type of generative model that can generate realistic data in many domains, such as images, language, and music. These models have been extensively studied in the last few years, resulting in a wide variety of applications, particularly in computer vision. However, there are several common challenges that must be solved to successfully apply GANs to real-world problems, such as being able to generate diverse images and ensure stable training. In the first part of this presentation I will introduce the theory that underpins the current understanding of why GANs work well, before describing the most common types of GAN model and some of their applications in computer vision. Environmental Data Science often deals with large amounts of 2D multi-channel data, such as satellite images or maps of climate variables. GANs are ideally suited for any application that involves generating this kind of data, and in the second part of this presentation I will summarise the work that has been done with GANs in this area.

A longer, more detailed lecture introducing GANs is available at https://www.youtube.com/watch?v=wFsI2WqUfdA.

Join Zoom Meeting https://zoom.us/j/6708259482?pwd=Qk03U3hxZWNJZUZpT2pVZnFtU2RRUT09

Meeting ID: 670 825 9482 Passcode: 9fkTAc

This talk is part of the AI4ER Seminar Series series.

This talk is included in these lists:

• AI4ER Seminar Series
• Cambridge Centre for Data-Driven Discovery (C2D3)
• Cambridge talks
• Chris Davis' list
• Interested Talks
• Trust & Technology Initiative - interesting events
• bld31
• https://zoom.us/j/6708259482?pwd=Qk03U3hxZWNJZUZpT2pVZnFtU2RRUT09
• ndk22's list
• ob366-ai4er
• rp587

Note that ex-directory lists are not shown.