Copy
View this email in your browser
Climate Change AI Newsletter
Our five-day virtual workshop on Tackling Climate Change with Machine learning may have drawn to a close, but we hope that the ideas explored there will continue to intrigue and inspire in the months to come. This newsletter links to resources from the workshop, in addition to the usual compilation of calls for papers, funding opportunities, courses, and competitions of general interest. Read on to learn how aerial imagery can be used to verify carbon offsets, how bioacoustics can support ecosystem conservation, and how machine learning can inform low carbon mobility.

Do you have opportunities to share or other content you would like to see included in the newsletter? Get in touch at info@climatechange.ai.

Workshop Highlights

Thank you to everyone who participated in last week’s virtual workshop! We were delighted to provide a venue for authors to share their research, for speakers to share their perspectives, and for participants to engage with one another. All talks from the main workshop are available at this link. The full live stream is available here. Selected talks from virtual booths are available here.

Our four keynote speakers. From top left, clockwise: Ciira wa Maina, explaining data collection for ecosystem monitoring projects, Georgina Campbell Flatter, describing the importance of reliable weather data for increasing robustness to floods, Stefano Ermon sharing work on remote sensing for food security, and Dan Morris directing machine learning practitioners to starting points for projects.
 

Here are a few highlights from the week,

  • Ciira wa Maina (Dedan Kimathi University of Technology) discussed the development of bioacoustic monitoring systems, from initial conversations with collaborators, to the design and placement of novel sensors, to the analysis of derived spectrograms.
  • Georgina Campbell Flatter (ClimaCell) discussed efforts to improve access to weather data globally, and the importance of such data in agriculture and transportation, among other domains.
  • Stefano Ermon (Stanford University) described strategies to overcome the challenge of limited labeled data in remote sensing, leveraging auxiliary labels from Wikipedia and self-supervision, and framed these advances within the broader field of computational sustainability.
  • Dan Morris (AI for Earth) explained the deep connections between land use, biodiversity, and climate change, shared a wealth of resources for students and practitioners hoping to enter the field, and of course included many pictures of animals.
  • Our fireside chat with Catherine Nakalembe (University of Maryland) highlighted the use of remote sensing to adapt to shifting landscapes, illustrating how experiences in the field and lab can shape institutional decisions and individual lives.
  • Cathy Wu (MIT), Konstantin Klemmer (University of Warwick), and Prithvi Acharya (Carnegie Mellon) discussed how machine learning can support low carbon mobility, drawing from reinforcement learning and graph neural networks, for example.
  • Climate Change AI members presented two tutorials, a Machine Learning 101 and a Climate Change 101, to explain the most important concepts (and address the most common misunderstandings) in each field.
If you found the workshop interesting, we encourage you to continue the conversation on our discussion forum or Twitter (@ClimateChangeAI).
  Calls for Submissions

Cities as Complex Systems: With their highly centralized services, cities are ideal sites for decarbonization. The Public Library of Science is inviting studies analyzing cities from a complex systems perspective in an issue edited by Marta Gonzalez (UC Berkeley) and Diego Rybski (Potsdam Institute for Climate Impact Research). Papers due July 3.

Machine Learning for Earth Observation Data (MACLEAN): As explored in the remote sensing fireside chat on Cross-Cutting Methods day of the workshop, earth observation data are relevant for a variety of climate mitigation and adaptation projects. The MACLEAN workshop at ECML / PKDD 2020, which will be held in Ghent, Belgium in September, will provide a venue for machine learning research focused on remote sensing applications. Submissions due June 9.
 
  Competitions & Courses
 
ProjectX, an initiative out of the University of Toronto’s Artificial Intelligence Group, is hosting a research competition (forum) on climate change and AI. The challenge is open to undergraduate students at any university, and runs from September to October. Team registration closes on May 31.

The
GeoCLEF competition pairs remote sensing data with localized species counts, with the goal of advancing methodology for biodiversity and habitat monitoring. The challenge is live and accepting submissions. Read the accompanying paper for background on the data and starting points for analysis.

The Forecasts within Energy Markets International
Summer School, which runs from September 21 - 25 in Annweiler am Trifels, Germany, will provide students with a foundation in both the political and economic theory and the algorithmic techniques needed to do work at the forefront of forecasting in energy markets. Registration is currently open, and interested students are encouraged to send a statement of interest and CV to the course’s organizers.

The edX course,
Climate Change: The Science and Global Impact, recently started, and will provide students with a foundation in the study of climate change. The course is taught by Michael Mann (Penn State), an author of the original “hockey stick” graph.
 
   Funding & Fellowships
 
The US Department of Energy has released two funding opportunities (press release, forum). The first, on AI and Decision Support for Complex Systems invites proposals on the use of AI to advance decision-making in scientific computing. The call closes on July 5. The second, on Scientific Machine Learning for Modeling and Simulations will fund proposals related to machine learning for acceleration of costly simulations and closes on May 29.

The AIMS Next Einstein Initiative Fellowship Program for Women in Climate Change Science has opened applications for its 2020 cohort. The fellowship supports work by women scientists contributing to a sustainable societal response to climate change. Fellows will be hosted by an African institution and are expected to develop projects that advance climate change mitigation or adaptation. Applications due June 30.
  Readings
Stephan Rasp, Soukayna Mouatadid, Peter Dueben, Sebastian Scher, Jonathan Weyn, Nils Thuerey

(Best Paper Award) Medium-range weather forecasting is more challenging than either short- or long-term forecasting. This is because long-term forecasting tends to be concerned with climatic averages, while short-term forecasts tend to be amenable to direct extrapolation. The authors of this work prepare a machine learning-ready dataset, along with a collection of competitive baselines, designed to help benchmark progress on this task. The project is an example of the type of data curation and open science effort that can advance the conversation on climate change and machine learning.

TrueBranch: Metric Learning-based Verification of Forest Conservation Projects
Simona Santamaria , David Dao, Björn Lütjens, Ce Zhang

(Best Proposal Award) Carbon offsetting efforts are only as effective as the monitoring, reporting, and verification (MRV) protocols that accompany them. It is increasingly common to ask landowners who are receiving payments for ecosystem services to send aerial drone imagery of their property, to verify forest cover. However, these data are easy to fake in a way that artificially inflates the estimated carbon stock of a plot of land. The proposal in TrueBranch is to compare submitted drone imagery with publicly available, but lower resolution, satellite imagery, checking for suspicious discrepancies. The authors provide empirical evidence that their strategy can identify deliberately tampered drone imagery while controlling false positives, thus strengthening MRV protections.
   Jobs

Studentships
  • PhD positions in detection and quantification of the emissions of greenhouse gases and pollutants using satellite data [Institut Pierre Simon Laplace, near Paris]
  • PhD positions in aerosol-cloud interactions, through iMIRACLI [Institute of Data Science in Jena, Germany and University College London, United Kingdom, forum]
  • PhD positions in smart grids [University of Lincoln, United Kingdom]
Industry
Twitter
Website
LinkedIn
Facebook
Copyright © 2020 Climate Change AI, All rights reserved.


Dates and deadlines reported in this newsletter may change, and CCAI is not responsible for any inadvertent inaccuracies.

Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.

Icons made by Nhor Phai, Payungkead, and surang.

Email Marketing Powered by Mailchimp