All topics

Topic

Climate risk and geospatial machine learning

Climate work here runs from research into production. At IBM Research, machine learning models for long range seasonal forecasting of 2m temperature and precipitation were published at NeurIPS and EGU and deployed into IBM PAIRS Geoscope, a petabyte scale geospatial platform. In insurance, the same kind of thinking became geospatial flood and natural catastrophe risk models across more than 230,000 insured properties. The proposed doctoral research, not yet registered, continues the thread with physics informed self supervised learning for flood mapping from radar imagery.

2 work · 2 research · 2 publications · Wikidata Q3433166

Work

Research

Publications

Subjects that share work with this one. Those with their own page are linked.