Public Lectures

Machine Learning at LASP: How we train computers to identify patterns in space data

Published on September 21, 2020

Speaker:Wendy Carande
Date:Wednesday, Dec 02, 2020
Location:Zoom (Access video of this past lecture form this lecture's webpage. Click title above.)

Seminar Abstract:


Machine learning, a way of programming a computer to improve at a task without human intervention, has gained popularity in Earth and space sciences in recent years. This rise in prevalence is due to increasing data volumes as well as algorithmic advances. When you think about machine learning or artificial intelligence, you might imagine a robot with human-like mental capabilities à la Blade Runner. In reality, many scientists and software engineers use machine learning for less exotic tasks, such as teaching a computer to recognize anomalies in images or find patterns in vast swaths of data.

In this talk, LASP software engineer Wendy Carande will discuss how machine learning is applied to the data collected at LASP. From predicting when a large solar flare is likely to occur to identifying outliers in telemetry data, machine learning has become a valuable tool for many LASP projects, but it has also presented challenges that will be explored.​