[Esip-machinelearning] ESIP ML Cluster newsletter (July newsletter) & cluster during ESIP summer meeting
Yuhan Douglas Rao
yuhan.rao at gmail.com
Mon Jul 13 15:20:00 EDT 2020
Dear all,
I hope you are all ready and excited for the ESIP virtual summer meeting.
Our cluster members will be busy during the summer meeting. There are three
sessions organized by cluster members. In the July edition of our cluster
newsletter, you will find a preview of these sessions and we hope we will
see you in some of these sessions and other great discussions during the
summer meeting,
The newsletter also included exciting developments of the labeled dataset
for machine learning in Earth sciences as well as COVID-19 related
resources that may be relevant to the community. We hope you enjoy reading
the newsletter!
See you at the summer meeting!
Best,
Yuhan (Douglas) Rao
ESIP Community Fellow - Machine Learning
Earth Science Information Partners
Machine Learning Cluster Newsletter
July 13, 2020
------------------------------
Greetings from Cluster Chair and Fellow
We hope to see you at the ESIP 2020 Virtual Summer Meeting, starting this
week. Even as it’s ESIP’s first virtual meeting, the meeting attendance
list is the longest in ESIP history (with 501 registered participants)! ML
Cluster members have organized three sessions focusing on machine learning
and its applications, described below. Also, changes are forthcoming in
Cluster leadership.
>From outside the Cluster, more data is becoming available, such as the
Radiant Earth Foundation’s large-scale collection of labeled geospatial
data for land cover in ten European countries. We also provide references
to ‘plainspoken’ articles about machine learning and its applications in
Earth sciences.
Finally, resources useful for ML and COVID-19 applications in Earth Science
are being made available in the form of initiatives, solicitations, data,
and computing resources, we provide some references below.
If you have any news and topics related to machine learning in Earth
sciences, please feel free to let us know so we can share it with the
community.
Anne Wilson, Ronin Institute
Yuhan Rao, North Carolina Institute for Climate Studies
July 13, 2020
Inside the cluster
Highlights of ESIP-ML and cluster members’ activities
-
2020 Virtual ESIP Summer Meeting. The ESIP 2020 Summer meeting starts
this week, see the schedule here on Sched
<https://2020esipsummermeeting.sched.com>. The ML Cluster members have
organized three virtual sessions focused on machine learning and its
applications in Earth sciences.
-
The first session, Machine Learning Tutorials
<https://2020esipsummermeeting.sched.com/event/cIvp/machine-learning-tutorials>,
(July 16, 2:00 pm Eastern time) provides a brief introduction of machine
learning workflow for Earth science applications and three hands-on
tutorials developed through the support of 2019 Funding Friday
award. This
introductory session demonstrates 3 ML learning strategies and
is intended
for beginners interested in experimenting with machine learning for their
own Earth science applications.
-
The second session (July 21, 2:00 pm Eastern time) focuses on the
discussion of organizational strategies for adopting machine learning
in the Earth science domain
<https://2020esipsummermeeting.sched.com/event/cT76/organizational-strategies-standards-and-policies-for-ml-are-there-any>.
We have an exciting panel consisting of representatives from government
agencies (NOAA and USGS) as well as industries (Microsoft and Element 84)
to discuss organizational best practices and strategies for using ML.
This session is also a lead into the follow-up session, where
we’ll discuss
how the Cluster can move forward to best support ESIP members, such as
these organizations. If you have any questions related to either
session, please contact Anne Wilson <anne.wilson at ronininstitute.org>
and/or Yuhan Rao <yuhan.rao at gmail.com>.
-
The third session (July 22, 4:00 pm Eastern time) led by Ziheng
(Jensen) Sun will bring together practitioners of artificial
intelligence to showcase innovative applications of AI in data-driven
Earth sciences
<https://2020esipsummermeeting.sched.com/event/cIuH/understanding-and-utilizing-ai-in-data-driven-earth-science>.
The session aims to help the community accelerate the engagement
between AI
and Earth data and improve our ability to deliver value-added information
faster and more accurately.
-
Passing of the Chair torch. After two years in the role of Cluster
Chair, Anne is stepping down. Ziheng (Jensen) Sun has agreed to take the
helm through December 2020, which is also when Yuhan's fellowship with the
cluster ends. Discussion and planning for Cluster leadership starting in
January 2021 will be a topic for discussion within the Cluster this fall.
If you are interested to lead the next chapter of the cluster, please let
Anne and Jensen know. We would like to thank Anne for her leadership to the
cluster in the past two years!
All about Machine Learning
Information relevant to ESIP-ML community
-
Radiant Earth released the BigEarthNet Benchmark Archive. The Radiant
Earth Foundation recently released the BigEarthNet Benchmark Archive
<https://www.radiant.earth/?gclid=Cj0KCQjwuJz3BRDTARIsAMg-HxW9rs2N568ixNuExSMJFDW2tqtiOLJDMx_-Y7FGb-3Sgwl4_-kUgdYaAjJbEALw_wcB>.
It is a large-scale collection of labeled geospatial data for land cover in
ten European countries. The archive is made available via Radiant MLHub
<http://www.mlhub.earth/> for free access. This dataset collection “consists
of 590, 326 Sentinel-2 image patches with spectral bands at 10, 20, and
60-meter resolution. The satellite images were acquired in different
seasons between June 2017 and May 2018 over Austria, Belgium, Finland,
Ireland, Kosovo, Lithuania, Luxembourg, Portugal, Serbia, and Switzerland.
Each patch is annotated with multiple land cover labels documenting the
spatial distribution of every land cover class across the dataset region.”
To learn more about the archive, please visit Radiant Earth Foundation’s
original post
<https://medium.com/radiant-earth-insights/bigearthnet-benchmark-archive-now-available-on-radiant-mlhub-the-open-repository-for-geospatial-d6c5dbe898c4>
.
-
Lacuna Fund to support the development of labeled data for agriculture.
Machine learning models can only be as good as its training data. There
have been great initiatives in other fields to establish benchmark training
data for ML model development and evaluation. Such initiatives are still at
an early stage for Earth sciences. To close the data gap, Lacuna Fund
recently announced its first request for a proposal to support the “the
creation, expansion and maintenance of labeled data in the agriculture
space”. More detailed information for this opportunity can be found
on Lacuna
Fund website <https://lacunafund.org/agriculture/>.
-
National Geographic AI for Earth Innovation Grants. To address the many
pressing scientific questions and challenges facing our planet, we must
increase global understanding of how human activity is affecting natural
systems and create a community of change, driven by data and cutting-edge
technology. Modern technologies, such as satellite imaging, bioacoustic
monitoring, environmental DNA, and genomics, can capture data at a global
scale, but also produce massive, complex data sets. Artificial intelligence
(AI) and cloud computing can capitalize on the potential of such data,
leading to faster and more meaningful insights and creating the opportunity
for transformative solutions. Learn more here
<https://www.nationalgeographic.org/funding-opportunities/grants/what-we-fund/ai-earth-innovation/>.
The deadline for the RFP is July 22, 2020.
-
Selected recent plainspoken articles about machine learning and its
applications in Earth sciences.
-
Wheeling, K. (2020), Improving atmospheric forecasts with machine
learning, Eos, 101, https://doi.org/10.1029/2020EO145068. Published
on 02 June 2020.
-
Hudson, M. (2020), Core progress in AI has stalled in some
fields, Science,
368, DOI: 10.1126/science.368.6494.927
<https://science.sciencemag.org/content/368/6494/927?utm_campaign=toc_sci-mag_2020-05-28&et_rid=79608803&et_cid=3343139>.
Published on 29 May 2020.
-
Wheeling, K. (2020), Machine learning improves weather and climate
models, Eos, 101, https://doi.org/10.1029/2020EO142422. Published on
07 April 2020.
-
Sima, R. J. (2020), Combining AI and analog forecasting to predict
extreme weather, Eos, 101, https://doi.org/10.1029/2020EO140896.
Published on 04 March 2020.
-
Kelvey, J. (2020), New study hints at bespoke future of lightning
forecasting, Eos, 100, https://doi.org/10.1029/2020EO139829.
Published on 13 February 2020.
-
Resources for COVID-19, Earth sciences, and machine learning. As the
invisible game-changer, COVID-19 continues to rock the world, we surveyed
the literature looking for topics on machine learning (ML), the COVID-19
virus, and Earth science. With a portion of the $2 trillion CARES Act
allocated to research and development, ML and AI are being heavily applied
in health, biology, genomics, epidemiology, and logistics support, etc.
But the application of ML towards COVID-related issues within the context
of Earth science per se seems not yet as extensive. One report, “Satellite
images, phone data help guide pandemic aid in at-risk developing countries
<https://techxplore.com/news/2020-06-satellite-images-pandemic-aid-at-risk.html>”,
describes using satellite imagery to determine the wealth of a region, its
road degradation, and other proxies for the need to determine a region’s
likelihood to need aid during the pandemic.
However, resources useful for ML and COVID-19 applications in Earth Science
are being made available in the form of initiatives, solicitations, data,
and computing resources. We provide some of them here.
-
Initiatives & Solicitations
-
NASA, “Making Innovative Use of NASA Satellite Data to Address
Environmental, Economic, and/or Societal Impacts of the COVID-19 pandemic
<http://www.spaceref.com/news/viewsr.html?pid=53475>”
-
ESA contests for remote sensing experts, ML scientists, and the
public to submit ideas on using satellite data to assess COVID impacts
<https://phys.org/news/2020-04-covid-satellites.html>.
-
CARTO has grant money for small nonprofits, charities, and NGOs for
using spatial data and “Location Intelligence”.
-
DOD Newton Award for Transformative Ideas during the COVID-19 Pandemic
<https://www.grants.gov/web/grants/view-opportunity.html?oppId=326034>.
This award is presented to a single investigator or team of up to two
investigators that develops a “transformative idea” to resolve
challenges,
advance frontiers, and set new paradigms in areas of immense potential
benefit to DoD and the nation at large. ”
-
Data & Literature
-
ESRI has made available maps, datasets, and some applications
<https://coronavirus-resources.esri.com> for COVID.
-
C3.ai has created a COVID-19 data lake, providing free “unified,
analysis-ready COVID-19 data’. They provide a knowledge graph that can be
queried to understand and do some analysis on this space of datasets.
-
CORD 19 <https://www.semanticscholar.org/cord19> provides a free
corpus of scholarly articles about the virus.
-
COVID Scholar <https://www.covidscholar.org> is a COVID-19 literature
search using natural language processing.
-
Digital Science is making all COVID-19 related published articles
<https://www.dimensions.ai/news/dimensions-is-facilitating-access-to-covid-19-research/>
available for free.
-
EPA is expanding research on COVID in the environment through
its homeland
security research webinar series
<https://www.epa.gov/homeland-security-research/homeland-security-research-webinar-series>
.
-
Computing resources
-
NCAR and the COVID-19 High-Performance Computing Consortium is making
the NCAR-operated Cheyenne supercomputer available for
COVID-related work.
Click this news releas
<https://news.ucar.edu/132724/ncar-operated-supercomputer-join-national-covid-19-computing-consortium>e
for more information.
-
UNH has created a nice list of COVID-19 Funding Opportunities,
“highlighting
opportunities and resources particularly suited to UNH
<https://www.unh.edu/research/find-funding/covid-19-funding-opportunities-research-priorities-resources>
.”
------------------------------
You are receiving this newsletter because you are part of the ESIP Machine
Learning Cluster. If you have any news want to share or questions, please
contact Anne Wilson (anne.wilson *at* ronininstitute.org) and Yuhan Rao
(yuhan.rao *at* gmail.com).
ML cluster newsletter - 2020 July - final.pdf
<https://drive.google.com/file/d/1XE_osJKcqwybs1hyWnuZvGn6MoB38mFl/view?usp=drive_web>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.esipfed.org/pipermail/esip-machinelearning/attachments/20200713/1168438d/attachment-0001.htm>
More information about the Esip-MachineLearning
mailing list