Dates: August 9-11, 2021
Format: 100% virtual, open to public - available as zoom webinar and as YouTube live stream.
PLEASE VISIT OUR YOUTUBE CHANNEL FOR ALL RECORDED SESSIONS:
https://www.youtube.com/channel/UCMYTOjm4uAI3xKWGY7_7xKA
Supported by NSF Awards under the Harnessing the Data Revolution (HDR) Program:
Background Call for Posters Agenda Confirmed Speakers Organizers Inaugural Workshop Register HERE!
Quicklinks to session details: Opening Session (ML1) Weather and Climate Aquatic Sciences Hydrology Translational Biology Closing Session (ML2)
Other workshop links: Workshop Home Workshop Logistics Poster Sessions Workshop Booklet
The second Knowledge Guided Machine Learning Workshop (KGML2021) is a three day virtual workshop August 9-11, 2021.
This workshop builds on the recognition that current black-box machine learning (ML) models have met with limited success in many scientific domains. This is due to their large data requirements, inability to produce physically consistent results, and lack of generalizability to out-of-sample scenarios. Instead of a purely data-driven approach that ignores decades (sometimes centuries) of accumulated knowledge in the science domains, there is an increasing interest in Knowledge Guided Machine Learning (KGML) approaches to integrate such knowledge into the ML models. This workshop will discuss recent advances made in this emerging area as well as their applications in the context of four domains: freshwater science, hydrology, climate and weather, and translational biology.
The workshop will include invited talks and poster sessions by data scientists (researchers in data mining, machine learning, and statistics) and researchers from four application areas (aquatic sciences, hydrology, atmospheric science, and translational biology) to discuss challenges, opportunities, and early progress in bringing scientific knowledge to machine learning. In doing so, it aims to foster interdisciplinary collaborations and interactions among these communities. The expected outcomes are the identification and greater understanding of challenges and opportunities in knowledge guided machine learning.
Previous Inaugural Workshop: The inaugural workshop took place August 18-20, 2020 virtually over zoom. Over 1000 people registered from over 30 countries, and a variety of topics and vibrant discussions were recorded and are available to watch on the KGML YouTube channel: https://z.umn.edu/kgmlworkshopyoutube. The audience consisted of a wide range of disciplines - data science, aquatic science, hydrology, weather/climate, and translational biology, seismology, ecology, agriculture, etc. Please visit the inaugural workshop webpage for the agenda along with slides and recorded video (for speakers who provided permission), workshop synopsis, and a program booklet with talk titles/abstracts/speaker-bio: https://sites.google.com/umn.edu/kgml/workshop
These workshops are funded by the NSF's Harnessing the Data Revolution (HDR) program involving researchers from the University of Minnesota, University of Wisconsin, Penn State, Colorado State University, University of Virginia, University of Illinois Urbana-Champaign, Carnegie Mellon University, University of Chicago, University of Pittsburgh, and George Mason University.
Everyone is invited to share their research, including works in progress, that are related to the workshop themes (as described in the background above), by submitting short abstracts for review. Accepted abstracts will be invited to submit a poster that will be presented in one of the three poster sessions as outlined in the agenda below. Posters will be presented virtually via Gather.Town.
Submission deadlines and guidelines can be found here.
(All times are listed in Central Time, UTC -5)
8/9, 9:30-12:25 Session one: Opening Session (ML1)
SPEAKERS:
9:30-9:55 Vipin Kumar, University of Minnesota and NSF HDR PI: KGML2021 Workshop Introduction and Overview (Presentation Slides) (Presentation Video)
9:55- 10:25 Elizabeth Barnes, Colorado State University: Controlled Abstention Networks: neural networks that say "I Don't Know" to learn better (Presentation Slides) (Presentation Video)
10:25- 10:55 Jordan Read, United States Geological Survey: Advancing Water Prediction With Knowledge-Guided Machine Learning Partnerships: Perspectives from the U.S. Geological Survey (Presentation Slides) (Presentation Video)
10:55-11:10 BREAK
11:10-11:35 Xiaowei Jia, University of Pittsburgh: Physics-Guided Machine Learning for Model Initialization Using Physical Simulations (Presentation Video)
11:35- 12:00 Zhenong Jin, University of Minnesota: Knowledge guided machine learning for agroecosystem sustainability: applications to modeling N2O emission and ecohydrology (Presentation Slides) (Presentation Video)
12:00- 12:25 Keynote: Animashree Anandkumar, California Institute of Technology: Enabling Zero-Shot Generalization in AI4Science (Presentation Video)
8/9, 12:30- 1:25 Poster Session day one: ML + Weather/Climate
8/9, 1:30-4:25 Session two: Weather and Climate
SPEAKERS
1:30-1:50 Timothy Delsole, George Mason University: Overview of Knowledge-Guided Machine Learning for Weather and Climate (Presentation Slides) (Presentation Video)
1:50-2:15 Pierre Gentine, Columbia University: Hybrid modeling (physics plus machine learning) to improve prediction of the hydrological cycle (Presentation Slides) (Presentation Video)
2:15-2:40 Katie Dagon, National Center for Atmospheric Research: Machine learning-based feature detection to associate precipitation extremes with synoptic weather events (Presentation Slides) (Presentation Video)
2:40-3:05 Peter Dueben, European Centre for Medium-Range Weather Forecasts: Challenges when preparing machine learning tools for use in operational weather predictions (Presentation Slides) (Presentation Video)
3:05-3:10 BREAK
3:10-3:35 Antonios Mamalakis, Colorado State University: Assessing methods of explainable artificial intelligence (XAI) by using attribution benchmark datasets (Presentation Slides) (Presentation Video)
3:35-4:00 Laurie Trenary, George Mason University: Skillful statistical prediction of sub-seasonal temperature by training on dynamical model data (Presentation Slides) (Presentation Video)
4:00-4:25 Pedram Hassanzadeh, Rice University: Building physical consistencies into neural networks for weather/climate modeling (Presentation Slides) (Presentation Video)
8/10, 9:30-12:25 Session three: Hydrology
SPEAKERS:
9:30-9:40 Opening by Session Moderators, Chris Duffy, Pennsylvania State University and John Nieber, University of Minnesota (Presentation Slides) (Presentation Video)
9:40-10:05 Yi Zheng, Southern University of Science and Technology, China: Uncovering flooding mechanisms through interpretive deep learning (Presentation Slides)
10:05-10:30 Andrew Bennett, University of Washington: Embedding neural networks to simulate turbulent heat fluxes in a process-based hydrologic modeling framework (Presentation Slides) (Presentation Video)
10:30-11:00 Ankush Khandelwal and Xiang Li, University of Minnesota: Source Aware Modulation for leveraging limited data from heterogeneous sources (Presentation Slides) (Presentation Video)
11:00-11:05 BREAK
11:05-11:30 Daniel Althoff, Stockholm University: Explainable machine learning: a peek into black-box models (Presentation Slides) (Presentation Video)
11:30-11:55 Katherine Ransom, United States Geological Survey: Process-informed machine learning predictions of nitrate in groundwater used for drinking supply in the conterminous United States (Presentation Slides) (Presentation Video)
11:55-12:20 Jonghyun Harry Lee, University of Hawaii: ML-based scalable data assimilation with hydrological applications (Presentation Slides) (Presentation Video)
8/10, 12:30- 1:25 Poster Session day two: Hydrology + Aquatic Sciences
8/10, 1:30-4:25 Session four: Aquatic Sciences
SPEAKERS:
1:30-1:50 Alison Appling, United States Geological Survey and Paul Hanson University of Wisconsin, Madison: Knowledge-guided machine learning on the rise in the aquatic sciences (Presentation Slides) (Presentation Video)
1:50-2:10 Charuleka Varadharajan, Lawrence Berkeley National Lab: Using Multi-scale Machine Learning Models to Develop a Predictive Understanding of the Impacts of Disturbances on River Water Quality (Presentation Slides) (Presentation Video)
2:10-2:30 Sam Oliver, United States Geological Survey: Process guided deep learning for decision-ready predictions (Recorded Presentation)
2:30-2:50 Janet Barclay, United States Geological Survey: Remember where you are: teaching a stream temperature model to embrace long-term groundwater exchange patterns (Presentation Slides) (Recorded Presentation)
2:50-3:00 BREAK
3:00-3:20 Lake Expedition 2020: a virtual collaboration of early career researchers integrating machine learning and aquatic sciences (Presentation Slides) (Presentation Video)
3:20-3:40 Moritz Feigl, University of Natural Resources and Life Sciences, Vienna, Austria: Learning from mistakes - Assessing the performance and uncertainty in process-based models (Presentation Slides) (Presentation Video)
3:40-4:00 Anuj Karpatne, Virginia Polytechnic Institute and State University: Biology-guided Neural Networks: Integrating Biological Knowledge with Neural Networks for Discovering Phenotypic Traits from Fish Images (Presentation Slides) (Presentation Video)
4:00-4:25 Aquatic Sciences Facilitated Discussion (Panel Video)
8/11, 9:30-12:25 Session five: Translational Biology
SPEAKERS:
9:30-9:50: Aidong Zhang, University of Virginia: Introduction to the knowledge transfer in translational biology (Presentation Slides) (Presentation Video)
9:50-10:30: Session Keynote, Russell Schwartz, Carnegie Mellon University: Learning how somatic mutability shapes cancer progression risk (Presentation Slides) (Presentation Video)
10:30-10:55: Tamer Kahveci, University of Florida: Counting motifs on evolving network topologies (Presentation Video)
10:55-11:10 BREAK
11:10-11:35 Jisoo Park, Novartis: Decoding cancer cell maps to guide precision medic (Presentation Slides) (Presentation Video)
11:35-12:00: Peter Kasson, University of Virginia: Better learning through chemistry: knowledge-guided inference of biomolecular kinetics (Presentation Video)
12:00-12:25: Brian Hie, Stanford University: Predicting evolution with neural language models (Presentation Slides) (Presentation Video)
8/11, 12:30- 1:25 Poster Session day three: Translational Biology + ML
8/11, 1:30-4:25 Session six: Closing Session (ML2)
SPEAKERS:
1:30-2:05 Keynote: George Karniadakis, Brown University: Approximating functions, functionals, and operators using deep neural networks for diverse applications (Presentation Video)
2:05-2:30 J Nathan Kutz, University of Washington: Data-driven model discovery and physics-informed learning (Presentation Slides) (Presentation Video)
2:30-2:55 Auroop Ganguly, Pacific Northwest National Laboratory, Northeastern University: Advancing the science of hydroclimatology and preparedness to flooding with integrated natural-build-human process models and data-driven sciences (Presentation Slides) (Presentation Video)
2:55-3:10 BREAK
3:10-3:35 Yan Liu, University of Southern California: Differential Graph Neural Networks for Physics-Informed AI Models (Presentation Video)
3:35-4:00 Henry Adams, Colorado State University: Topology in Machine Learning (Presentation Slides) (Presentation Video)
4:00-4:25 Arindam Banerjee, University of Illinois Urbana-Champaign: Learning for long range temporal prediction (Presentation Slides) (Presentation Video)
4:25 Closing Remarks
University of Minnesota: Vipin Kumar, John Nieber, Michael Steinbach, Ju Sun
University of Wisconsin-Madison: Hilary Dugan, Paul Hanson, Robert D. Nowak, Stephen Wright
Colorado State University: Elizabeth Barnes, Imme Ebert-Uphoff
University of Virginia: Kevin Janes, Aidong Zhang
George Mason University: Benjamin Cash, Timothy DelSole
United States Geological Survey (USGS): Alison Appling
University of Illinois, Urbana-Champaign: Arindam Banerjee
Pennsylvania State University: Chris Duffy
University of Chicago: Rebecca Willett
University of Pittsburgh: Xiaowei Jia
Carnegie Mellon University: Pradeep Ravikumar
Above: QR code to access slido via your mobile device
Alternatively, you can access slido via your browser: https://app.sli.do/event/uxi2q3kf
Event code #52107
(This section is for archival purposes only, as the workshop has ended, please go to our YouTube channel to see recorded presentations: https://www.youtube.com/channel/UCMYTOjm4uAI3xKWGY7_7xKA)
Please see our Workshop Logistics page for details on virtually attending, posing questions, and viewing posters.
The link to join the workshop as an attendee is (this link is the same for ALL SESSIONS):
Zoom: https://z.umn.edu/kgmlworkshop21
Youtube: https://z.umn.edu/kgmlworkshopyoutube21
Please note: joining via zoom is limited to 1000 attendees, therefore, attendees will be automatically redirected to the youtube livestream if we have reached the zoom maximum.
Q&A: You will be able to pose questions to presenters via slido, details include the QR code and caption to the left. If you have any questions about slido, please visit the workshop logistics page.
Posters: All poster sessions will be accessed via Gather.Town using this link: https://gather.town/app/UGsc1qY7AOLKDEJP/KGML2021 (Note this link is password protected and should be in your registration confirmation or attendee email. If you need the password, email kgmlworkshop@umn.edu to request it.)
Quicklinks to session details: Opening Session (ML1) Weather and Climate Aquatic Sciences Hydrology Translational Biology Closing Session (ML2)
Other workshop links: Workshop Home Workshop Logistics Poster Sessions