
Find out how AWS for Aerospace and Satellite is working with the Taylor Geospatial Institute on the Generative AI for Geospatial Challenge.
Loading summary
Maria Varmazas
Welcome to AWS in Orbit. I'm Maria Varmazas. We're working with AWS to bring you an in depth look at the transformative intersection of cloud computing, space technologies and generative AI. On AWS in Orbit, we're exploring not just what's possible, but what's meaningful in the realm of space and cloud innovation. We grapple with complex challenges and unparalleled opportunities that arise when we use space to address pressing issues right here on Earth. This is AWS in Orbit, the Generative AI for Geospatial Challenge. Today I'm joined by two amazing people. Nadine Alame from the Taylor Geospatial Institute and Salem El Nimri of AWS to talk talk about a fascinating joint venture that's kicking off right now to use generative AI and geospatial data to imagine big possibilities and take on the Earth's toughest problems. Today this venture is called the Generative AI for Geospatial Challenge. And here's more about it.
Nadine Alame
Hello Maria. I'm Nadine Alemy, the executive director of the Taylor Geospatial Institute. I like to think of myself as a cheerleader for all things geospatial. So location and mapping and climate data and disaster data and population dynamics, it's all geospatial. I've seen it from my days at NASA. I've seen it as a startup from an aviation information management and weather perspective. I've seen it from the nonprofit side as the CEO of the Open Geospatial Consortium. And now I'm trying to pull it all together with AWS to, I call it, accelerate the impact of this geospatial technology and data and tools and community into the world.
Maria Varmazas
Well, Salem, without further ado, over to you to do a, you gotta follow that act.
Salem El Nimri
Well, hi, my name is Salem Nimby. I'm basically part of that team at aws and we look at everything that has to do with aerospace and satellites, whether it is on the satellite communication side or the Earth observation side. And we have a lot of customers who are, who range all the way from the ones who are capturing image to the ones who are inferring intelligence out of this image. I lead the team of subject matter experts. My background have been with NASA for quite some time. I remember we were like brainstorming and Nadine and I, we were sitting together and we were like AI and generative AI and there is a big wave. We were talking and Nadine was like, why don't we try to tap into the Community of experts. The technology is here, and there is a lot of advancements on the AWS side that goes from our CHIPS developments to all our partners and the capabilities they are developing. And we wanted to bring the community of subject matter experts to leverage those services and capabilities so that we accelerate and we explore the art of the possible, whether it is for national defense or for the private sector and industries. And that's what we're excited about.
Maria Varmazas
So we got a bit of a background about this challenge, the genesis of this. Tell me a bit more about what it is and maybe a bit about the timeline. Let's start with just the basics of this first.
Nadine Alame
We kicked off this challenge on October 29th in St. Louis with Salem and the AWS team and the Taylor Geospatial Institute and the community. And I started that day emphasizing to people that they need to remember October 29, 2024, because this is when something unique was happening, right? It's this merging of a lot of data from space. Think about it, right? We get data from so many satellites. You know, it's not just imagery. It's also radar. It's hyperspectral. I mean, you name it, combine it with all the data that we have, you know, from sensors on Earth, right? The whole Internet of things. And the only way, the only way you can make sense of it today is artificial intelligence, right? And the only way you can actually do this is if you have this infrastructure that can scale with you, because nobody can just do this on their own. So we said, okay, like Salem said, let's unleash the community, let's unleash the world, right? So who's an entrepreneur, who's a startup, who's a researcher, who's somebody who's working in any type of business, even government, they're all welcome. We had them all during that day, and we gave them one month. One month. So October 29th to November 2029.
Maria Varmazas
Wow.
Nadine Alame
We're going to be getting exactly these proposals. Tell us your ideas. We ask people like, you know, what should we be doing here? This is powerful. And they said, would it be amazing if together in this challenge we can figure out how do we. I'm going to use it against Salem. Accelerate the analysis of deforestation and natural disasters. We had somebody from government actually saying something that hit. Can we do some analysis at global scale to anticipate flooding events like the one we just had in North Carolina?
Maria Varmazas
Right? Yeah.
Nadine Alame
This is what we're talking about. A lot of data, a lot of experts, expertise. What can you world do with this. That's the challenge.
Salem El Nimri
This is where we're after to make sense of all of this data. You have either two pairs of eyes that's looking at an image, and if you have thousands of images, it's going to be impossible to look through all of these. And that's where the power of AI and machine learning comes into play. All of this can be automated and we can learn more things than we used to before. And with generative AI and the introductions of large language models, we're going to bring some ease to how are we interacting. It's like we're speaking to that data and getting it to tell us what's in there.
Nadine Alame
Of course, what I heard several times during that day, and maybe that resonates with the audience, is think about it as ChatGPT, but for Earth, for anything that happens on Earth. Right. You know, when should I evacuate my family if a hurricane strikes? Right. What should I plant in my farm, in my land, given the climate change patterns, Right. It's like. But you say it like Salem is saying, you interact with the data and you don't have to be the specialist scientist. That's the innovation. That's what we're looking for.
Salem El Nimri
Simplify, simplify and aid the government, because the government, as they are trying to do a lot of planning and management, this is all aid for the government. They'll be able to do more with less resources, basically.
Maria Varmazas
That's fantastic. I am so fascinated by what you both have described as this very special moment in time where we have this intersection of all this technological capability that just wasn't really that possible before, without a lot of manual work, that we are at this very special time where so much can be accelerated. I guess I'm going to use that word too, by AI. You've both sort of described these fascinating ideas. I'm curious, are there any that you're particularly excited about that you're hoping to see pitches for when all is revealed pretty soon?
Nadine Alame
So I think there's a spectrum. I'm rooting for. The other theme that came up during the day, again, the power of this is this theme of building blocks. So the innovations could be one building block, but us looking at all these innovations can put the building blocks to create this. ChatGPT for Earth, ChatGPT for climate change, ChatGPT for agriculture. Right. I'm looking for the little things. Right. That are building blocks. Say we had actually a couple talks during that day. You know how when we talk about AI, everybody thinks about the identification of Cats, right? You have pictures of cats, right? But guess what? It's not as simple with a flood. It's not as simple with a landslide. It's not as simple with a hurricane. Because these things move and they're, they're more complex scientifically. Right. So I'm looking for even sort of like the feature extraction we call it of these phenomena. Right. This detection, this monitoring of this phenomena. Right. Because that's one building block.
Maria Varmazas
Salem, same question to you. What are you sort of rooting for? What are you hoping to see?
Salem El Nimri
I'm hoping to see how we are able to leverage the private sector to help our governments and bring more to become more informed as they are making the decisions that impacts all of our lives. There is an influx of data that's coming in from the private sector going into space. We see that in the rate of launches that's happening. And there's a lot of this split between Earth observation and satellite communications, to say the least between all of these. And I think the governments will be able to leverage all of this and learn and get more intelligence into managing their day to day business, managing security, managing infrastructure and planning for the future. One of the biggest challenges that you see when it comes to Earth observation and specifically remote sensing is that each instrument is actually specific and their data format, structure and what it can and cannot do. And if you try to look at all of this, each data point is important, but you're going to spend a lot of time translating. It would be awesome if we can leverage AI ML and generative AI to look at all this data. And instead of us spending more time in formatting the data to try to say, okay, Now I spent 60, 70% of my time in the formatting and now I'm ready to do analytics, to just dive into the analytics or dive into the outcomes. So you will speak and start looking for these patterns. And that will take some tedious work out of the way. That's the whole idea. Our scientists and researchers, they need to be able to focus on these things than to focus on the tedious things. And if we get to a point where we can normalize all of this, we will have people from even outside those industries come and bring their brain power and say, you know what, let me see what I can do with this. And you really don't know what's the art of the possible here? And that's what we're trying to do with this challenge is scratching the art of the possible.
Maria Varmazas
That's amazing. Given that the deadline is a month for the ideas. How baked do these ideas need to be? Are these very conceptual?
Nadine Alame
At this point, I think it's all over. That's why I think the target audience is startups and entrepreneurs as well as researchers. Because usually these people have a million ideas in their head. And what we're doing here is just making it possible for them to actually come in and experiment. Right. So AWS is providing a million dollars worth of credits, we're providing support for geospatial. It's just come in because part of this is also understanding what are the challenges. And I think in the process is a learning experience for everybody. Right. And what we're aiming for in this case is the art of the possible. Let's find those ideas. And that's why I go back to the building blocks that came up during the kickoff, because accelerating could be building one building block for everybody to use so they can, like Celine was saying, you go from 80% to 70% to eventually just 20% on preparation to actually solve your problem.
Salem El Nimri
Preparations. I think it takes the majority of any researcher or scientist time just to prep this data, put it in the format that is readable and usable, and then you need to bring another source of data and another source of data because you know, the more sources of data you will have, the more you are informed and the more you can extract out of this data. And this is what I'm looking for is how can we, how can we simplify? And I will use this, accelerate all of this so that we can reach to a quicker, more informed decision.
Maria Varmazas
The next steps basically are end of November is the deadline for ideas and then as you mentioned, four months till. Just walk me through it all again, if you don't mind.
Nadine Alame
So the submission of ideas November 29th and then we'll get back to people by the end of the year. The plan is from January to April, the creative, amazing, smart people will have four months. And we are planning to get together in April, May, possibly right around the Geoint Symposium to actually showcase it. I also want to say so, which I think goes back to how big the idea should be. And this is a short timeframe. Team up with people who can help make your idea become real or prove your idea so that you can take it to the next level. So we're expecting teaming up of academia, research, industry, large, small, we had very excellent representation of governments in the kickoff. Team up so that eventually I think this is the shortest path to solution. Right, because you have the team already.
Salem El Nimri
Collaboration is key because you need collaborations from the data providers, to the data scientists, to the people who think outside the box and say, you know what, it's amazing. This is change detection. Okay, what can we do with it? You can go anywhere from monitoring a site to detecting flooding, to clearing highways from debris, or managing power lines, cutting some of the canopies that's out there. There's no limit to what you can do. And I agree with unity in teams, collaboration is key to help push this through.
Maria Varmazas
Well, I'm really looking forward to hearing what the submissions are. I'm really excited about it. It's just such a fascinating challenge and the thought of that we're building towards an Earth GPT just blows my mind. I'm sure the possibilities there really are paradigm shifting in all sorts of ways. It's just going to be absolutely incredible to see as we build towards that future. One of my wrap up question here, maybe what kind of people would you like to see involved that maybe wouldn't normally be involved in this? People who you want to get involved who might not normally think of a challenge like this as being their realm, I suppose. What's their call to action? Who do you want to see?
Nadine Alame
What came up during the kickoff day and I was happily surprised when several speakers were talking about start with the user, start with the community or the agency or the group that has a problem. Usually it's like oh gen AI for geospatial is technical, but the agriculture problem is not technical. The disaster response preparedness is not technical. The patterns of just human migration because of climate change are not technical. So I would love to see the users actually represented so that we're actually solving real problems. We're not imagining problems. The people with the problems are also coming with the people who can help them.
Salem El Nimri
And I can't agree more with you Nadine. I would love to see those users and they come and say you know what, I have this problem and how can we solve it? And it could be as well. I will touch on health and being able to monitor from space and anticipate if there is an outbreak of some sort. Yes, we cannot detect the virus, but we can detect the environmental changes or the landscape changes where they happen. And if we see anything that is out of the norm, then those are early signs for us to be more proactive instead of reactive. Usually in most of the things even like disaster response. It's like a disaster has to happen and then you respond to it. I really look forward to the point where we can be more proactive and prevent those things from happening and if it is out of our ability to prevent, at least mitigate and reduce the damage that could result out of these things. Yes, we can't control Mother Nature, but we can prepare for what may come our way and preserve life and property and all of these things.
Maria Varmazas
Yeah, being able to anticipate, that's a very powerful thought, how amazing that would be. And it's amazing that we're actually talking about this as a real possibility. It's just incredible. This has been a fascinating discussion. As I said, I'm really looking forward to seeing, and I'm sure you both are too, what comes out of this. So any thoughts you want to leave our audience with? Nadine, I'll start with you.
Nadine Alame
I want to leave them with November 29th on our website, which we will share with the podcast. This is not just a fun exercise. We need this kind of coming together, collaborating across domains, across communities to figure out how we can remain on this earth, how our planet remains sustainable for our kids. So I will leave with. While this is fun and cutting edge, it's also urgent and that's why how you started this podcast. This coming together now is amazing in so many ways. So I'm looking forward to redefining collaboration across sectors, using cloud, using space, using all the data, you know, for all of us in this challenge. So I also want to thank AWS for providing the infrastructure and the community that essentially they serve. It's great. Thank you.
Salem El Nimri
Thank you. And what I'm looking for is I'm looking for all the people to collaborate and join. No idea is too small or too big. And we are here collaborating with TGI to basically tap into the best of the best out there and say, come over, let us help you and we'll get to the next step together. It's all collaboration from our side providing infrastructure chips and training modules to the other side on tgi, providing the subject matter experts and the community around it. This is only the beginning. We are going to come back and we look forward to come back on this show with some of our customers and winners to tell us about the exciting journey they had throughout this challenge, what innovations they were able to come up with and bring to the table, and how we were able to influence them. So we are lucky and privileged to have this going on and there is no idea that is too small or too big. Just please come over, apply and we look forward to collaborating with everyone. Thank you. And thank you, Nadine, for this opportunity and your amazing team.
Nadine Alame
And Maria, thank you. Back at you. 100%.
Maria Varmazas
And that's it for AWS in Orbit, the Generative AI for Geospatial Challenge. A special thanks to Nadine Alame and Salem El Nimri for joining us today. For additional resources from this episode and for more episodes in the AWS in Orbit series, check out our show notes@space.n2k.com AWS this episode was produced by Alice Carruth and powered by aws. Our AWS producer is Laura Barber. Our associate producer is Liz Stokes. We're mixed by Elliot Peltzman and Trey Hester with original music by Elliot Peltzman. Our executive producer is Jennifer Iban. Our executive editor is Brandon Karpf. Simone Petrella is our president. Peter Kilpy is our publisher, and I'm your host, Maria Varmazas. Thanks for listening. We'll see you next time.
Nadine Alame
Sa.
Summary of "AWS in Orbit: The Generative AI for Geospatial Challenge"
Podcast: T-Minus Space Daily
Host: Maria Varmazas
Guests: Nadine Alame (Executive Director, Taylor Geospatial Institute), Salem El Nimri (AWS Aerospace and Satellite Specialist)
Release Date: November 23, 2024
In the November 23, 2024 episode of AWS in Orbit, hosted by Maria Varmazas, the spotlight is on the Generative AI for Geospatial Challenge, a collaborative initiative between AWS and the Taylor Geospatial Institute. This challenge aims to harness the transformative power of cloud computing, space technologies, and generative AI to tackle some of Earth's most pressing issues through advanced geospatial data analysis.
Nadine Alame, the Executive Director of the Taylor Geospatial Institute, describes herself as a "cheerleader for all things geospatial," emphasizing her extensive background spanning NASA, startups in aviation information management, and her role as CEO of the Open Geospatial Consortium. Her collaboration with Salem El Nimri from AWS seeks to "accelerate the impact of this geospatial technology" by integrating AWS's scalable infrastructure with global geospatial data and expert communities.
Salem El Nimri elaborates on his role within AWS, where he leads a team focused on aerospace and satellite technologies, encompassing both satellite communication and Earth observation. Drawing from his NASA experience, Salem highlights the necessity of leveraging AI to manage the vast influx of geospatial data, facilitating advancements across national defense, private sectors, and various industries. He states, "We wanted to bring the community of subject matter experts to leverage those services and capabilities so that we can accelerate and explore the art of the possible" ([02:56]).
The Generative AI for Geospatial Challenge was officially launched on October 29, 2024, in St. Louis, as Nadine underscores the significance of this date—marking the convergence of diverse data streams from multiple satellites and Earth sensors. This includes imagery, radar, hyperspectral data, and the broader Internet of Things (IoT). Nadine emphasizes, "The only way you can make sense of it today is artificial intelligence" ([04:31]).
Participants from all backgrounds—entrepreneurs, startups, researchers, government entities—are encouraged to submit proposals within a one-month timeframe, culminating on November 29, 2024. The challenge invites innovative ideas that leverage geospatial data and generative AI to address complex problems such as deforestation, natural disaster analysis, and global flood anticipation.
The challenge seeks solutions that can accelerate the analysis of deforestation and natural disasters, as well as anticipate flooding events on a global scale, akin to recent incidents like the North Carolina floods ([06:42]). Nadine likens the envisioned platform to "ChatGPT for Earth," enabling users to interact with geospatial data conversationally. She elaborates, "It's like when should I evacuate my family if a hurricane strikes... What should I plant in my farm, given the climate change patterns" ([07:36]).
Salem adds that generative AI and large language models will revolutionize data interaction, allowing users to "speak to the data and get it to tell us what's in there" ([06:53]). This approach simplifies complex geospatial analysis, making it accessible to non-specialists and aiding governments in planning and resource management.
Managing the vast and varied geospatial data requires sophisticated AI and machine learning techniques. Salem points out that traditional methods of manually analyzing images are impractical given the volume, highlighting the necessity of automation: "All of this can be automated and we can learn more things than we used to before" ([07:36]).
Nadine focuses on the technical challenges, such as detecting dynamic and complex phenomena like floods and landslides, which are far more intricate than identifying static objects like cats in images. She emphasizes the need for "feature extraction" as fundamental building blocks for effective geospatial analysis ([09:22]).
The challenge is open to a wide array of participants, including startups, entrepreneurs, and researchers. Nadine encourages those with "a million ideas in their head" to experiment and engage, supported by AWS's provision of $1 million in credits and geospatial support ([13:19]). The initiative seeks not only technical solutions but also real-world applications driven by users who understand the problems firsthand.
Salem echoes this sentiment, advocating for collaboration across sectors to harness the full potential of incoming geospatial data from the private sector. He envisions a platform where varying data sources are normalized, allowing researchers to "focus on the analytics or dive into the outcomes" rather than data formatting ([10:30]).
Nadine emphasizes the importance of timely collaboration, stating, "Team up with people who can help make your idea become real or prove your idea so that you can take it to the next level" ([15:17]). Salem adds that effective collaboration among data providers, scientists, and innovative thinkers is crucial for pushing the boundaries of what's possible ([16:21]).
A key theme of the challenge is the creation of foundational building blocks that can be utilized across various applications. This modular approach enables the development of comprehensive solutions like Earth GPT, an envisioned platform that integrates multiple data sources and AI capabilities to provide actionable insights for diverse sectors.
The ultimate goal is to develop a system akin to "Earth GPT," which would enable proactive management of environmental and societal challenges. By anticipating events such as natural disasters or health outbreaks, this platform aims to shift from reactive responses to preventative measures. Salem envisions applications like monitoring environmental changes to foresee health crises, thereby allowing for early intervention and mitigation ([18:46]).
As the episode wraps up, Nadine and Salem reiterate the urgency and excitement surrounding the Generative AI for Geospatial Challenge. Nadine urges potential participants to remember the submission deadline and highlights the critical nature of collaborative efforts for Earth's sustainability: "While this is fun and cutting edge, it's also urgent" ([20:27]).
Salem emphasizes inclusivity, stating, "No idea is too small or too big. Please come over, apply, and we look forward to collaborating with everyone" ([21:32]). The initiative promises to redefine collaboration across sectors, leveraging cloud and space technologies to address real-world problems effectively.
This episode of AWS in Orbit underscores a pivotal moment where advanced AI, robust cloud infrastructure, and extensive geospatial data converge to unlock unprecedented solutions for Earth's challenges. The Generative AI for Geospatial Challenge stands as a testament to the power of collaborative innovation, inviting diverse minds to contribute to a sustainable and intelligent future.