Timezone: »

AI for Earth Sciences
Surya Karthik Mukkavilli · Karthik Kashinath · Johanna Hansen · Kelly Kochanski · Tom Beucler · Mayur Mudigonda · Paul D Miller · Chad Frischmann · Amy McGovern · Pierre Gentine · Gregory Dudek · Aaron Courville · Daniel Kammen · Vipin Kumar

Sat Dec 12 06:45 AM -- 09:00 PM (PST) @ None
Event URL: https://ai4earthscience.github.io/neurips-2020-workshop/ »

Our workshop proposal AI for Earth sciences seeks to bring cutting edge geoscientific and planetary challenges to the fore for the machine learning and deep learning communities. We seek machine learning interest from major areas encompassed by Earth sciences which include, atmospheric physics, hydrologic sciences, cryosphere science, oceanography, geology, planetary sciences, space weather, volcanism, seismology, geo-health (i.e. water, land, air pollution, environmental epidemics), biosphere, and biogeosciences. We also seek interest in AI applied to energy for renewable energy meteorology, thermodynamics and heat transfer problems. We call for papers demonstrating novel machine learning techniques in remote sensing for meteorology and geosciences, generative Earth system modeling, and transfer learning from geophysics and numerical simulations and uncertainty in Earth science learning representations. We also seek theoretical developments in interpretable machine learning in meteorology and geoscientific models, hybrid models with Earth science knowledge guided machine learning, representation learning from graphs and manifolds in spatiotemporal models and dimensionality reduction in Earth sciences. In addition, we seek Earth science applications from vision, robotics, multi-agent systems and reinforcement learning. New labelled benchmark datasets and generative visualizations of the Earth are also of particular interest. A new area of interest is in integrated assessment models and human-centered AI for Earth.

AI4Earth Areas of Interest:
- Atmospheric Science
- Hydro and Cryospheres
- Solid Earth
- Theoretical Advances
- Remote Sensing
- Energy in the Earth system
- Extreme weather & climate
- Geo-health
- Biosphere & Biogeosciences
- Planetary sciences
- Benchmark datasets
- People-Earth

Sat 6:45 a.m. - 6:55 a.m.

AI for Earth Sciences, Workshop Founder & Chair, S. Karthik Mukkavilli

Karthik Mukkavilli
Sat 7:00 a.m. - 7:02 a.m.

Sensors and Sampling, Session Chair, Johanna Hansen

Johanna Hansen
Sat 7:02 a.m. - 7:27 a.m.

WARPLab's research focuses on both the science and systems of exploration robots in extreme, communication starved environments such as the deep sea. It aims to develop robotics and machine learning-based techniques to enable search, discovery, and mapping of natural phenomena that are difficult to observe and study due to various physical and information-theoretic challenges.

WARPLab is headed by Yogesh Girdhar, and is part of the Deep Submergence Laboratory (DSL), and the Applied Ocean Physics & Engineering (AOPE) department at Woods Hole Oceanographic Institution.

Yogesh A Girdhar
Sat 7:30 a.m. - 7:43 a.m.

Talk Title: "Eyes in the sky without boots on the ground: Using satellites and machine learning to monitor agriculture and food security during COVID-19"

Hannah Kerner is an Assistant Research Professor at the University of Maryland, College Park. Her research focuses on developing machine learning solutions for remote sensing applications in agricultural monitoring, food security, and Earth/planetary science. She is the Machine Learning Lead and U.S. Domestic Co-Lead for NASA Harvest, NASA’s food security initiative run out of the University of Maryland.

Hannah Kerner
Sat 7:43 a.m. - 8:06 a.m.

Talk Title: Autonomous Robot Manipulation for Planetary Science: Mars Sample Return, Climbing Lava Tubes

This talk will highlight work at NASA on robotic missions from a machine vision perspective. The discussion will focus on the science questions that NASA hopes to answer through returned samples from Mars and the challenges imposed on robotic systems used for scientific data collection.

Related Papers: http://renaud-detry.net/publications/Pham-2020-AEROCONF.pdf https://www.liebertpub.com/doi/10.1089/ast.2019.2177

Renaud Detry is the group leader for the Perception Systems group at NASA's Jet Propulsion Laboratory (JPL). Detry earned his Master's and Ph.D. degrees in computer engineering and robot learning from ULiege in 2006 and 2010. He served as a postdoc at KTH and ULiege between 2011 and 2015, before joining the Robotics and Mobility Section at JPL in 2016. His research interests are perception and learning for manipulation, robot grasping, and mobility, for terrestrial and planetary applications. At JPL, Detry leads the machine-vision team of the Mars Sample Return surface mission, and he leads and contributes to a variety of research projects related to industrial robot manipulation, orbital image understanding, in-space assembly, and autonomous wheeled or legged mobility for Mars, Europa, and Enceladus.

Renaud Detry
Sat 8:07 a.m. - 8:14 a.m.

Visual analysis of complex fish habitats is an important step towards sustainable fisheries for human consumption and environmental protection. Deep Learning methods have shown great promise for scene analysis when trained on large-scale datasets. However, current datasets for fish analysis tend to focus on the classification task within constrained, plain environments which do not capture the complexity of underwater fish habitats. To address this limitation, we present DeepFish as a benchmark suite with a large-scale dataset to train and test methods for several computer vision tasks. The dataset consists of approximately 40 thousand images collected underwater from 20 habitats in the marine-environments of tropical Australia. The dataset originally contained only classification labels. Thus, we collected point-level and segmentation labels to have a more comprehensive fish analysis benchmark. These labels enable models to learn to automatically monitor fish count, identify their locations, and estimate their sizes. Our experiments provide an in-depth analysis of the dataset characteristics, and the performance evaluation of several state-of-the-art approaches based on our benchmark. Although models pre-trained on ImageNet have successfully performed on this benchmark, there is still room for improvement. Therefore, this benchmark serves as a testbed to motivate further development in this challenging domain of underwater computer vision.

Alzayat Saleh, Issam Hadj Laradji, David Vázquez
Sat 8:14 a.m. - 8:23 a.m.

Forestry is a major industry in many parts of the world, yet this potential domain of application area has been overlooked by the robotics community. For instance, forest inventory, a cornerstone of efficient and sustainable forestry, is still traditionally performed manually by qualified professionals. The lack of automation in this particular task, consisting chiefly of measuring tree attributes, limits its speed, and, therefore, the area that can be economically covered. To this effect, we propose to use recent advancements in three‐dimensional mapping approaches in forests to automatically measure tree diameters from mobile robot observations. While previous studies showed the potential for such technology, they lacked a rigorous analysis of diameter estimation methods in challenging and large‐scale forest environments. Here, we validated multiple diameter estimation methods, including two novel ones, in a new publicly‐available dataset which includes four different forest sites, 11 trajectories, totaling 1458 tree observations, and 14,000 m2. From our extensive validation, we concluded that our mapping method is usable in the context of automated forest inventory, with our best diameter estimation method yielding a root mean square error of 3.45 cm for our whole dataset and 2.04 cm in ideal conditions consisting of mature forest with well‐spaced trees. Furthermore, we release this dataset to the public (https://norlab.ulaval.ca/research/montmorencydataset), to spur further research in robotic forest inventories. Finally, stemming from this large‐scale experiment, we provide recommendations for future deployments of mobile robots in a forestry context.

Jean-François is a Ph.D. student at McGill’s Mobile Robotics Lab, under the supervision of prof. Dave Meger. He is interested in model-based RL for mobile robot navigation in unstructured environments such as forests, tundra or underwater. Previously he was a masters student at the Northern Robotics Laboratory (Norlab), working on lidar mapping and perception for forestry applications.

Jean-François Tremblay
Sat 8:30 a.m. - 8:55 a.m.

Moderated by Johanna Hansen

Johanna Hansen, Yogesh A Girdhar, Hannah Kerner, Renaud Detry
Sat 8:55 a.m. - 9:00 a.m.

Ecology, Session Chair, Natasha Dudek

Natasha Dudek
Sat 9:00 a.m. - 9:25 a.m.

Program Director of Microsoft AI for Earth

D. Morris
Sat 9:25 a.m. - 9:55 a.m.

Talk Title (tentative): ML and control of parasitic diseases of poverty in tropical and subtropical countries, with a special focus on schistosomiasis Professor at Stanford University Senior Fellow at Stanford Woods Institute for the Environment

Giulio De Leo
Sat 9:55 a.m. - 10:05 a.m.
Graph Learning for Inverse Landscape Genetics (Regular Talk - Ecology Session) Video
Prathamesh Dharangutte
Sat 10:05 a.m. - 10:15 a.m.
Segmentation of Soil Degradation Sites in Swiss Alpine Grasslands with Deep Learning (Regular Talk - Ecology Session) Video
Maxim Samarin
Sat 10:15 a.m. - 10:20 a.m.
Novel application of Convolutional Neural Networks for the meta-modeling of large-scale spatial data (Lightning Talk - Ecology Session) Video
Kiri A. Stern
Sat 10:20 a.m. - 10:25 a.m.
Understanding Climate Impacts on Vegetation with Gaussian Processes in Granger Causality (Lightning Talk - Ecology Session) Video
Miguel Morata Dolz
Sat 10:25 a.m. - 10:30 a.m.
Interpreting the Impact of Weather on Crop Yield Using Attention (Lightning Talk - Ecology Session) Video
Tryambak Gangopadhyay
Sat 10:30 a.m. - 10:55 a.m.

Moderated by Natasha Dudek

Natasha Dudek, D. Morris, Giulio De Leo
Sat 10:55 a.m. - 11:00 a.m.

By S. Karthik Mukkavilli

Karthik Mukkavilli
Sat 11:00 a.m. - 11:25 a.m.
Pierre Gentine (Water - Session Keynote) Video Pierre Gentine
Sat 11:25 a.m. - 11:40 a.m.

Long Oral (15m)

Martin Gauch
Sat 11:40 a.m. - 11:55 a.m.

Long Talk (15m)

Grey S Nearing
Sat 11:55 a.m. - 12:10 p.m.

Long Talk (15m)

Gonzalo Mateo-García
Sat 12:10 p.m. - 12:20 p.m.
Efficient Reservoir Management through Deep Reinforcement Learning (Regular Talk - Water Session) Video
Xinrun Wang
Sat 12:20 p.m. - 12:45 p.m.

Moderated by S. Karthik Mukkavilli

Karthik Mukkavilli, Pierre Gentine, Grey S Nearing
Sat 12:45 p.m. - 1:15 p.m.

Prof Milind Tambe

Director, Center for Research on Computation & Society Gordon McKay Professor of Computer Science Harvard John A. Paulson School of Engineering and Applied Sciences Mail: Maxwell Dworkin 125, 33 Oxford Street, Cambridge, MA 02138

Director for AI for Social Good Google India Research Center


Milind Tambe
Sat 1:15 p.m. - 1:25 p.m.
Q/A and Discussion
Karthik Mukkavilli, Mayur Mudigonda, Milind Tambe
Sat 1:25 p.m. - 1:30 p.m.

By Tom Beucler

Tom Beucler
Sat 1:30 p.m. - 1:55 p.m.
Michael Pritchard (Atmosphere - Session Keynote) Video Mike Pritchard
Sat 1:55 p.m. - 2:20 p.m.

Identifying Opportunities for Skillful Weather Prediction with Interpretable Neural Networks

Elizabeth A. Barnes
Sat 2:20 p.m. - 2:35 p.m.

Long Talk (15m)

Lukas Kapp-Schwoerer
Sat 2:35 p.m. - 2:50 p.m.

Long Talk (15m)

Nadia Ahmed
Sat 2:50 p.m. - 2:55 p.m.
Towards Data-Driven Physics-Informed Global Precipitation Forecasting from Satellite Imagery (Lightning Talk - Atmosphere Session) Video
Valentina Zantedeschi, Valentina Zantedeschi
Sat 2:55 p.m. - 3:25 p.m.
Q/A and Discussion for Atmosphere Session (Q/A and Discussion)
Tom Beucler, Mike Pritchard, Elizabeth A. Barnes
Sat 3:25 p.m. - 3:30 p.m.

By Karthik Kashinath

Karthik Kashinath
Sat 3:30 p.m. - 3:55 p.m.
Stephan Mandt (Simulations, Physics-guided, and ML Theory - Session Keynote) Video Stephan Mandt
Sat 3:55 p.m. - 4:20 p.m.
Rose Yu (Simulations, Physics-guided, and ML Theory - Session Keynote) Video Rose Yu
Sat 4:20 p.m. - 4:30 p.m.
Generating Synthetic Multispectral Satellite Imagery from Sentinel-2 (Regular Talk - ML Theory) Video
Hamed Alemohammad
Sat 4:30 p.m. - 4:40 p.m.
Multiresolution Tensor Learning for Efficient and Interpretable Spatiotemporal Analysis (Regular Talk - ML Theory) Video
Raechel Walker
Sat 4:40 p.m. - 4:50 p.m.
Climate-StyleGAN : Modeling Turbulent ClimateDynamics Using Style-GAN (Regular Talk - ML Theory) Video
Rishabh Gupta
Sat 4:50 p.m. - 4:55 p.m.
Interpretable Deep Generative Spatio-Temporal Point Processes (Lightning Talk - ML Theory) Video
Shixiang Zhu
Sat 4:55 p.m. - 5:00 p.m.
Completing physics-based model by learning hidden dynamics through data assimilation (Lightning Talk - ML Theory Session) Video
Arthur Filoche
Sat 5:00 p.m. - 5:20 p.m.

Moderated by Karthik Kashinath and Mayur Mudigonda

Karthik Kashinath, Mayur Mudigonda, Stephan Mandt, Rose Yu
Sat 5:20 p.m. - 5:25 p.m.

By Mayur Mudigonda

Mayur Mudigonda
Sat 5:25 p.m. - 6:00 p.m.
Q/A and Panel Discussion for People-Earth with Dan Kammen and Milind Tambe (Q/A and Panel Discussion)
Mayur Mudigonda, Daniel Kammen, Milind Tambe
Sat 6:00 p.m. - 6:05 p.m.

By Kelly Kochanski

Kelly Kochanski
Sat 6:05 p.m. - 6:20 p.m.
Soft Attention Convolutional Neural Networks for Rare Event Detection in Sequences (Spotlight Talk - Solid Earth Session) Video
Mandar Kulkarni
Sat 6:20 p.m. - 6:30 p.m.
An End-to-End Earthquake Monitoring Method for Joint Earthquake Detection and Association using Deep Learning (Regular Talk - Solid Earth) Video
Weiqiang Zhu
Sat 6:30 p.m. - 6:40 p.m.
Single-Station Earthquake Location Using Deep Neural Networks (Regular Talk - Solid Earth) Video
Charles Mousavi
Sat 6:40 p.m. - 6:45 p.m.
Framework for automatic globally optimal well log correlation (Lightning Talk - Solid Earth Session) Video
Oleh Datskiv
Sat 6:45 p.m. - 7:00 p.m.
Q/A and Discussion for Solid Earth (Q/A and Discussion)
Kelly Kochanski
Sat 7:00 p.m. - 7:05 p.m.

By Karthik Kashinath

Karthik Kashinath
Sat 7:05 p.m. - 7:30 p.m.
Stephan Rasp (Benchmark Datasets - Session Keynote) Video Stephan Rasp
Sat 7:30 p.m. - 7:45 p.m.

Long Talk (15m)

Catherine Tong
Sat 7:45 p.m. - 8:00 p.m.

Long Talk (15m)

Samriddhi Singla, Tianhui Diao
Sat 8:00 p.m. - 8:10 p.m.
LandCoverNet: A global benchmark land cover classification training dataset (Regular Talk - Benchmark Datasets Session) Video
Hamed Alemohammad
Sat 8:10 p.m. - 8:20 p.m.
Applying Machine Learning to Crowd-sourced Data from Earthquake Detective (Regular Talk - Benchmark Datasets Session) Video
Omkar Ranadive
Sat 8:20 p.m. - 8:25 p.m.
An Active Learning Pipeline to Detect Hurricane Washover in Post-Storm Aerial Images (Lightning Talk - Benchmark Datasets Session) Video
Evan Goldstein
Sat 8:25 p.m. - 8:30 p.m.
Developing High Quality Training Samples for Deep Learning Based Local Climate Classification in Korea (Lightning Talk - Benchmark Datasets Session) Video
Minho Kim
Sat 8:30 p.m. - 8:55 p.m.
Q/A and Discussion for Benchmark Datasets (Q/A and Discussion)
Karthik Kashinath
Sat 8:55 p.m. - 9:00 p.m.
Workshop Closing Remarks
Karthik Mukkavilli
Sat 8:55 p.m. - 8:55 p.m.
Posters (Posters - Break (On Demand Pre-recorded Not Livestreamed)) Video
Karthik Mukkavilli
Sat 9:00 p.m. - 9:00 p.m.
Bias correction of global climate model using machine learning algorithms to determine meteorological variables in different tropical climates of Indonesia (Poster - Atmosphere Session) [ Video ] Video
Juan Nathaniel
Sat 9:00 p.m. - 9:00 p.m.
Optimising Placement of Pollution Sensors in Windy Environments (Poster - Atmosphere Session) [ Video ] Video
Sigrid Passano Hellan
Sat 9:00 p.m. - 9:00 p.m.
Temporally Weighting Machine Learning Models for High-Impact Severe Hail Prediction (Poster - Atmosphere Session) [ Video ] Video
Amanda Burke
Sat 9:00 p.m. - 9:00 p.m.
Integrating data assimilation with structurally equivariant spatial transformers: Physically consistent data-driven models for weather forecasting (Poster - Atmosphere Session) [ Video ] Video
Ashesh Chattopadhyay
Sat 9:00 p.m. - 9:00 p.m.
Unsupervised Regionalization of Particle-resolved Aerosol Mixing State Indices on the Global Scale (Poster - Atmosphere Session) [ Video ] Video
Zhonghua Zheng
Sat 9:00 p.m. - 9:00 p.m.
MonarchNet: Differentiating Monarch Butterflies from Those with Similar Appearances (Poster - Benchmark Datasets Session) [ Video ] Video
Thomas Chen
Sat 9:00 p.m. - 9:00 p.m.
Nowcasting Solar Irradiance Over Oahu (Poster - Atmosphere Session) [ Video ] Video
Peter Sadowski
Sat 9:00 p.m. - 9:00 p.m.
Semantic Segmentation of Medium-Resolution Satellite Imagery using Conditional Generative Adversarial Networks (Poster - ML Theory) [ Video ] Video
Hamed Alemohammad
Sat 9:00 p.m. - 9:00 p.m.
Spectral Unmixing With Multinomial Mixture Kernel and Wasserstein Generative Adversarial Loss (Poster - Sensing and Sampling Session) [ Video ] Video
Savas Ozkan
Sat 9:00 p.m. - 9:00 p.m.
Interpretability in Convolutional Neural Networks for Building Damage Classification in Satellite Imagery (Poster - Sensing and Sampling Session) [ Video ] Video
Thomas Chen
Sat 9:00 p.m. - 9:00 p.m.
Towards Automated Satellite Conjunction Management with Bayesian Deep Learning (Poster - Sensing and Sampling Session) [ Video ] Video
Francesco Pinto
Sat 9:00 p.m. - 9:00 p.m.
Domain Adaptive Shake-shake Residual Network for Corn Disease Recognition (Poster - Sensing and Sampling Session) [ Video ] Video
Yuan Fang
Sat 9:00 p.m. - 9:00 p.m.
A Comparison of Data-Driven Models for Predicting Stream Water Temperature (Poster - Water Session) [ Video ] Video
Helen Weierbach
Sat 9:00 p.m. - 9:00 p.m.
Inductive Predictions of Extreme Hydrologic Events in The Wabash River Watershed (Poster - Water Session) [ Video ] Video
Nicholas Majeske
Sat 9:00 p.m. - 9:00 p.m.
Predicting Streamflow By Using BiLSTM with Attention from heterogeneous spatiotemporal remote sensing products (Poster - Water Session) [ Video ] Video
Udit Bhatia - IITGN

Author Information

Karthik Mukkavilli (University of California, Irvine, Berkeley Lab & McGill)
Karthik Kashinath (LBNL)
Johanna Hansen (McGill University)
Kelly Kochanski (University of Colorado Boulder)

Earth science researcher using machine learning to make better predictions about natural hazards and climate change.

Tom Beucler (University of California, Irvine)
Mayur Mudigonda (UC Berkeley)
Paul D Miller (DJ Spooky)
Chad Frischmann (Drawdown)
Amy McGovern (University of Oklahoma)
Pierre Gentine (Columbia University)
Gregory Dudek (McGill University & Samsung Research)
Aaron Courville (U. Montreal)
Daniel Kammen (University of California, Berkeley)
Vipin Kumar (University of Minnesota)

More from the Same Authors