CS&E Wins 2022 SDM Best Paper Award

SDM is a top-tier conference in the data science field that selects a small number of full-length submitted papers after a stringent review process. Out of all accepted papers, the committee awards one best paper that significantly advances the state-of-the-art methods in data science. The best paper award recognizes outstanding work appearing at the annual conference, and winners are picked by a committee delegated by the program chairs. 

The winning work by Tayal and team addresses the problem of predicting the behavior of environmental systems, such as lakes, under a changing climate. Specifically, the paper presents a novel machine learning methodology for improving the prediction of lake temperatures that have direct impact on water quality and its suitability for sustaining aquatic life. Existing methods for modeling lake temperature only perform well for bodies of water for which sufficient amounts of high-quality observations are available. These models often fail to generalize to other, less observed scenarios. 

Additionally, lake temperatures are impacted not just by climate variables (such as temperature, wind speed, precipitation), but also by their inherent characteristics such as depth and clarity of the lake. For many lakes, these characteristics are not known, making it harder to predict the temperature profiles under different climate scenarios. Tayal’s paper addresses both of these issues by a novel approach that uses invertible neural networks along with state of the art deep-learning methods to better predict temperature profiles, as well as missing static features.

“This work highlights the important role machine learning can play in dealing with environmental problems”, said Kumar. “Advances such as these are made possible by a highly productive collaboration our team has had with freshwater scientists at USGS over the past five years and state of the art computing infrastructure available at the Minnesota Supercomputing Institute.”

Please join CS&E in congratulating the entire team on receiving this award.