Timezone: »

 
Tutorial
Climate Change: Challenges for Machine Learning
Arindam Banerjee · Claire Monteleoni

Mon Dec 08 06:30 AM -- 08:30 AM (PST) @ Level 2, Room 210 a, b
Event URL: http://research.microsoft.com/apps/video/?id=238888 »

Despite the scientific consensus on climate change, drastic uncertainties remain. Crucial questions about changes in regional climate, trends of extreme events such as heat waves, heavy precipitation, and mega-storms, and understanding how climate varied in the distant past, must be answered in order to improve predictions, assess impacts and vulnerability, and aid mitigation and adaptation efforts. Machine learning can help answer such questions and shed light on climate change. Similar to the case of bioinformatics, the study of climate change provides a data-rich scientific domain in which cutting-edge tools from machine learning can make a major impact. Further, such questions give rise to new challenges for the design of machine learning algorithms.

This tutorial will give an overview of impactful open questions about climate change, highlight recent successes of machine learning in this domain, and outline significant remaining challenges. Machine learning problems in climate change include prediction, reconstruction, causal attribution, analysis of quantiles and extremes, and exploratory data analysis. Challenges arise because the climate system is extremely complex, comprised of physical processes and their interactions, and the data is massive, high-dimensional, and spatiotemporal, with non-stationarity and potential long-range dependencies over time and space.

Author Information

Arindam Banerjee (University of Minnesota, Twin Cities)

Arindam Banerjee is a Professor at the Department of Computer & Engineering and a Resident Fellow at the Institute on the Environment at the University of Minnesota, Twin Cities. His research interests are in machine learning, data mining, and applications in complex real-world problems in different areas including climate science, ecology, recommendation systems, text analysis, and finance. He has won several awards, including the NSF CAREER award (2010), the IBM Faculty Award (2013), and six best paper awards in top-tier conferences.

Claire Monteleoni (University of Colorado Boulder)

Claire Monteleoni is an associate professor of Computer Science at University of Colorado Boulder. Previously, she was an associate professor at George Washington University, and research faculty at the Center for Computational Learning Systems, at Columbia University. She did a postdoc in Computer Science and Engineering at the University of California, San Diego, and completed her PhD and Masters in Computer Science, at MIT. She holds a Bachelors in Earth and Planetary Sciences from Harvard. Her research focuses on machine learning algorithms and theory for problems including learning from data streams, learning from raw (unlabeled) data, learning from private data, and climate informatics: accelerating discovery in climate science with machine learning. Her work on climate informatics received the Best Application Paper Award at NASA CIDU 2010. In 2011, she co-founded the International Workshop on Climate Informatics, which is now in its fourth year, attracting climate scientists and data scientists from over 14 countries and 26 states.

More from the Same Authors