- This event has passed.
GEOAI Roundtable
March 21, 2023 @ 10:00 am - 1:00 pm
OGC invites location, AI, special interest, and the broader technology communities that foresee potential use of geospatial data in their work to attend our GeoAI Working Group’s Roundtable event. We want to hear from the location community about the exciting GeoAI work you all are doing, where GeoAI’s future is headed, and what data and interoperability challenges need to be addressed along the way.
What is GeoAI?
Geospatial artificial intelligence (GeoAI) is an emerging scientific discipline and an amalgamation of artificial intelligence (AI) with spatial computing to better understand the physical world around us at the levels of an individual, communities, cities, nations, and the planet using geospatial data. GeoAI is a highly interdisciplinary field that uses computing, geography, spatial data science, statistics, and engineering ideas. The unique features of geospatial data like autocorrelation, geometric nature, and multiple modalities require the AI algorithms to be developed in a fundamentally different way, making GeoAI a separate but symbiotic discipline with AI.
Why Does GeoAI Matter?
AI and Machine Learning have found their way into almost every sector of society, with important and far-reaching consequences. With both Geospatial and AI technology evolving at an incredible pace, the time to plan, discuss, and deliver the future of Geo AI is now. Therefore discussing the reach, and challenges associated with this is a debate that many organizations are having.rnrnGeoAI matters to the OGC from two perspectives – firstly that more and more member-provided solutions, pilots and applications across the organization are fueled by it, and secondly, that applicable standards need to be developed for realizing the full potential of GeoAI.
Roundtable Themes
u003c!u002du002d wp:paragraph {u0022fontSizeu0022:u0022mediumu0022} u002du002du003ernrnThe OGC GeoAI Working Group conducted a survey last year. It highlighted that there are a range of cross-cutting themes and challenges which deserve a more in-depth discussion. Based on the responses received and in-line with OGC’s purpose, the first roundtable will focus on following cross-cutting technical themes:rnrnu003c!u002du002d /wp:paragraph u002du002du003e u003c!u002du002d wp:list {u0022orderedu0022:true} u002du002du003ernu003colu003ern tu003cli style=u0022list-style-type: noneu0022u003ernu003colu003eu003c!u002du002d wp:list-item u002du002du003ern tu003cliu003eu003cstrongu003eTraining data sourcing/availability: u003c/strongu003eLack of availability of good training datasets is the most significant concern that the GeoAI community highlights as a challenge. It involves collecting ground truth information through various means like ground sensors or crowdsourcing.u003c/liu003ernu003c/olu003ernu003c/liu003ernu003c/olu003ernu003c!u002du002d /wp:list-item u002du002du003e u003c!u002du002d wp:list-item u002du002du003ernu003colu003ern tu003cli style=u0022list-style-type: noneu0022u003ernu003colu003ern tu003cliu003eu003cstrongu003eModel u0026amp; data drift: u003c/strongu003eThe concept of drift refers to the degradation of an AI model’s performance due to changes in data and relationships between input and output variables. Such changes occur over space and time. Being able to handle drift is vital for increasing the real-world adoption of GeoAI.u003c/liu003ernu003c/olu003ernu003c/liu003ernu003c/olu003ernu003c!u002du002d /wp:list-item u002du002du003e u003c!u002du002d wp:list-item u002du002du003ernu003colu003ern tu003cli style=u0022list-style-type: noneu0022u003ernu003colu003ern tu003cliu003eu003cstrongu003eInteroperability challenges: u003c/strongu003eCurrently, GeoAI work has been happening in silos within different companies for specific applications, therefore no standards are there for data formats, metadata, annotations, or communication among various GeoAI microservices. Without standards an integrated and seamless ecosystem of applications can not. Even measuring performance and reporting results as per standard definitions would give this market a big boost.u003c/liu003ernu003c/olu003ernu003c/liu003ernu003c/olu003ernu003c!u002du002d /wp:list-item u002du002du003ernrnu003c!u002du002d /wp:list u002du002du003e u003c!u002du002d wp:paragraph u002du002du003ernrnAs part of their talk, we would therefore like lightning talk guests to engage with these key questions – how can GeoAI applications fuel the OGC’s important work over the coming years, and how canrnrnu003c!u002du002d /wp:paragraph u002du002du003e
Roundtable Agenda
Introductions: GeoAI – Setting the Stage
| Nadine Alameh | OGC | About the OGC (5 minutes) |
| Kyoungsook KIM | GeoAI DWG | About the GeoAI DWG (5 minutes) |
| Lokendra Chauhan | GeoAI DWG | GeoAI Survey Results (10 minutes) |
Lightning Talks – Interoperability Theme
| Samantha Arundel | USGS | How good is good enough? Knowing when research is ready for operation (7 minutes) |
| Ryan Avery | DevelopmentSeed | Making GeoAI Models Searchable and Consumable (7 minutes) |
| Jimmy Chou | Feng Chia University | Advancements in Flood Alert Systems by Geo-AI: The Role of Streaming Information and Deep Learning (7 minutes) |
| Ankit Jha | IIT Bombay | GAF-Net: Improving the Performance of Remote Sensing Image Fusion using Novel Global Self and Cross Attention Learning (7 minutes) |
| Ajay Gupta | HSR Health | Leveraging GeoAI to Advance Global Public Health (7 minutes) |
| Dan Pilone | Element 84 | Improving Interoperability and Speed in the Labeling and Training Process (7 minutes) |
| Jo Walsh | Ordnance Survey, UK | A National Mapping Agency’s role in GeoAI? (7 minutes) |
| All presenters | Q&A for the Interoperability Presentations (10 minutes) |
Lightning Talks – Training Data Theme
| Chris Andrews | Rendered AI | Synthetic Data: Taking control of your computer vision training data (7 minutes) |
| Ankit Sharma | IIIT Kota | Geo-Personalized News (7 minutes) |
| Akshay Loya | GIS Kernel | Detecting Debris for Disaster management (7 minutes) |
| Pengyuan Liu | NUIST China | Towards Human-centric GeoAI: Leveraging Crowdsourced Data for Urban Sidewalk Planning (7 minutes) |
| Neeraj Gudipati | Qen Labs | Automated “Human in Loop” Data Annotations & Harmonization of Heterogeneous Data (7 minutes) |
| Michael Putz | Blackshark AI | AI Accelerated Rapid Sensor to insights (7 minutes) |
| Sacha Leprêtre | Presagis | AI as a crucial foundation for creating Large-Scale 3D Digital Twins – GeoAI Challenges (7 minutes) |
GeoAI Domain Working Group Discussion
| GeoAI DWG Members | OGC | Panel discussion and closing (30 minutes) |
Become a Speaker
Have you been working with GeoAI applications that could benefit from standardization efforts, or OGC standards that interface with GeoAI ? Or do you have real world experience in the major themes and challenges of the workshop and can provide an interesting perspective for future work?
We’d love to have you share what you’ve been up to with the broader community. Our priority is to hear about real-world implementation from as many different organizations as possible, so don’t feel like you need a ‘new innovation’. And of course it’s a great venue to share tools and spec extensions with others. Apply now to join the line-up of lightning speakers!
For the chance to speak at the GeoAI Roundtable event, complete this short survey and hit “Apply.”
Become a Sponsor
All sponsors will have their logos on the event website and in slides thanking sponsors during the event. Sponsors who commit before the general signup launch (tentatively on Feb 20, 2023) will have their logos included on the blog post announcing the signup. Sponsors who commit after that will get their logos included, but will miss out on the initial attention. Sponsors will receive recognition of their support throughout the event and its promotion. We firmly believe that the GeoAI movement is going to be a foundation of much future innovation, so we hope that by associating with GeoAI early your organization will get continued recognition in the future.
| TITLE | DESCRIPTION | COST |
| SUPPORTER | Supporters will be listed on the event website and blog posts, will be provided a personalized graphic showing support of the GeoAI roundtable, and will have their logo listed in recognition in keynote sessions. | $2000 USD |
| PARTNER | Partners will be listed with more prominent logos on the event website and blog posts. They will be thanked in the keynote sessions and have their logo included in tweets about the event. Partners will also receive a guaranteed lightning talk, with prominent placement in the session, on the subject of their choice. They will also be provided with a personalized graphic showing support of the GeoAI roundtable. | $8,000 USD |
| SUSTAINING PARTNER | Sustaining Partners will be listed with the most prominent logos on the event website and blog posts. They will be thanked directly by name in tweets about the event and in the keynote sessions. Sustaining Partners will also receive 1 guaranteed tutorial, as well as 1 special ‘double’ lightning talk of ten minutes, on the subject of their choice. They will also be provided with a personalized graphic showing support of the GeoAI roundtable. | $25,000 USD |
| INITIATIVE FOUNDER | We are also offering all the benefits of the sustaining partner level to those who commit early to substantial support of the follow-on GeoAI Initiative. This will directly fund a set of initiatives aimed at increasing the interoperability of geospatial data with the use of the GeoAI standards. If you are interested in being an initiative founder please get in touch and we can discuss what that will look like and include your organization in the early discussions that shape the initiative. | Custom |
About OGC GeoAI DWG
The OGC GeoAI Working Group is an open forum established by OGC to bring together the community of geoscientists, computer scientists, engineers, entrepreneurs, and decision makers from government, industry and academia to develop, share, and research the latest trends, successes, challenges, and opportunities in the field of AI with geospatial data. Themes covered by the working group to-date include: GeoAI in disaster responses, health care, smart energy, agriculture, manufacturing, self-driving vehicles, digital twins and more.