Exploring Graph Analytics in Electromagnetic Spectrum Management - Webinar with Expression Networks
Once again, welcome and thanks for being here today. Uh, in today's webinar, uh, we're going to take a look at graph analytics and electromagnetic spectrum management. Before I pass you off to our two speakers, uh, just a couple of practical notes first. You can drop your questions into the q and a panel at any time. We're saving some time at the end of the webinar to answer those. Second, you'll see some additional resources in the handouts panel, so free.
Feel free to take a look at those. Uh, I know there's a handout about, uh, expression, so feel free to share that, uh, with anyone who might be interested. And finally, this webinar is being recorded.
So we will be sharing the recording with you, uh, in the next couple of days, um, probably tomorrow in case you'd like to re-watch any part of it, uh, or share it with anyone who you think might be interested. And now I'm happy to pass you off to our two speakers today. Abir Ray and Jean Villedieu. Hello everyone. Uh, and welcome to today's webinar from Data to Dominance, graph analytics in the electomagnetic, spectrum.
Um, my name is, uh, Jean Villedieu. I'm one of the co-founders of Linkurious. Uh, and for today's webinar, I'm joined by Abir, uh, Ray from Expression Networks. I'm, uh, the CEO of expression.
Um. We, uh, mainly work, uh, in the government, federal government space in the us, uh, do a lot of work with the us DOD have found Linkurious to be a great partner to extend our, uh, base platform, uh, and to be able to provide, uh, graph analytics, uh, to areas of the military and other organizations within the US Federal government. So quick quarter about, uh, Linkurious. The company was founded in 2013.
Uh, we are a team of 50 plus persons, uh, with offices in France and in the us. And on a day, um, on a daily basis, more than 10,000 investigators, uh, throughout the world are using all products, uh, to make sense of complex 20 data, uh, and make smarter decisions. Uh, and here at Expression, uh, we've been around since 1997, just shy of 30 years. Uh, we, uh, have a data analytics platform that is utilized, as I mentioned, by the US Federal Government. Um, we also, uh, have, uh, quite a few, uh, data analytics and AI ML, uh, solutions that are deployed, uh, within the federal government.
We're based here in Washington, DC um, have been, uh, considered one of the fastest growing. Uh, small businesses over the last six, seven years. Employees, uh, and our expertise beyond our software platform, uh, range from, uh, software engineering to DevSecOps, uh, and, uh, you know, the full life cycle of, uh, the, the software, um, development stack. So today's agenda, uh, we are going to start with, uh, a brief intro about, uh, the Linkurious decision intelligence platform. Um, then we'll dive into the meat, uh, of today's webinar, talking about graph analytics, uh, in the electromagnetic spectrum.
Uh, and then we'll have some time, uh, at the end of the session for some QA. So question. So first, uh, Linkurious, what it is that we, um, that we, that we do, um, what we strongly believe in, uh, at Linkurious, is that context, uh, is really, really important, uh, to, uh, to make smart decisions. And you have a bit of an illustration of that here.
You have a seemingly normal client, uh, on the left, and maybe you can have, you know, 51 of the data points about that one particular person. Um, and these data points will not tell you that, uh, it's a, a bad client. But understanding the context of figuring out that this normal client is indirectly connected, uh, to, um, a potential criminal, um, the one person on the, on the right across, um, a production and address and, uh, and more that's, uh, going to help you figure out that there's a potential risk there and it can, um, make the difference between making the right or the wrong code.
Uh, and that's what we. We do that, uh, via software, uh, which is the, the Linkurious Decision Intelligence platform. Um, and it provides a very nice and, uh, and interactive UI for which analysts, uh, can conduct, uh, searches, um, explore the data, but also review alerts and all that in a collaborative and secure environment. Uh, and you can see what the software looks like, uh, on the, on the right. And here's a, a bit of an overview of of how it works. Uh, it all starts with data.
Um, clients typically have internal data, external data, um, and then within, um, the product, uh, we combine anti solution, the ability to, uh, to fuse different data sources. Um, the use of analytics, um, which help, um, detect suspicious, uh, patterns and enrich complex connected data. Uh, we also have the ability to, uh, to manage alerts, um, and, uh, provide some, uh, some features, um, already to case management.
Uh, and finally, uh, you can also search and conduct visual investigations, uh, via the, the link platform. Uh, at the end of the day. Uh, all of that add up, uh, to smarter and faster decisions, uh, for, for clients. So that's it for a brief introduction to, um, to ur.
Um, and now we'll learn that's the, the pattern of that to you. Thank you. Um, so, uh, as I mentioned, uh, the majority of our work is for the US feral government. Uh, we partnered with Linkurious over a year ago, uh, to be able to support one of our. Um, uh, Keystone products, uh, called EMDM, electromagnetic Battle Management, uh, within the US Federal Government, both on the civilian and then defense side.
Uh, expression provides the core platforms to manage, uh, spectrum, uh, through spectrum management, um, which is the management of, uh, spectrum licenses by federal agencies. Uh, use of spectrum, um, all the way to spectrum operations, uh, which is the core of our military product set. Um, and that allows, uh, the war fighter to dynamically understand, uh, what's available in the spectrum, um, where the spectrum maneuver space exists.
Uh, and then, uh, from there, how to navigate that maneuver space most effectively to, um, be able to affect, uh, their mission goals. Uh, just to give a little background since, you know, I've spent the last 25 years, uh, supporting this space, but a a lot of you might not know, uh, what the importance of electromagnetic spectrum is. Uh, if you're watching this on your cell phone, um, or, uh, you know, connected via wifi, uh, then you're using the EMS, you're using the electromagnetic spectrum, uh, and the, those wavelengths, uh, you know, are all the way down. Uh, like, and this is a, a chart of. The different use cases, uh, for the, you know, different parts of the electromagnetic spectrum, you know, all the way from am, uh, radio to the emissions, uh, within, uh, nuclear, uh, power reactor.
Um, everything you see around you, the, um, visually that you know, is part of the electromagnetic spectrum. Uh, the part that we deal in the most is communication. So how, uh, varied systems communicate with each other? Um, as you can imagine, uh, in the tactical edge, uh, for the DOD, uh, you know, everything that happens is, uh, utilizing the spec spectrum from radar to be able to, uh, you communicate and perform command and control for varied assets. Um, you know, it's, uh, it's the core, uh, of, uh, how the DOD, uh, operates. Uh, and without it, it would be impossible to do the things that, that we do. Um, being able to control the spectrum is very important.
So in the battlefield, uh, you know, those who are coming from Europe may have probably heard about GPS jamming. And so, you know, what, what is that? You know, GPS very essential to everything we do today. Uh, if you've ever used Google Maps or Apple Maps, um, getting from point A to point B. Um, you know, you don't think about it anymore, right? Um, you know, the, your phone knows exactly where you are and how to navigate you, uh, from point A to point B understands when there's congestion, uh, you know, in that route. Uh, and this is all, uh, possible because of GPS, which allows you to understand where you are and then, um, you know, wifi or cellular access.
To be able to communicate back, you know, where you are, back to some central system to be able to aggregate that information. So this works similarly for the DOD and you know, in, in the battlefield, uh, EMS is that central nervous system for military operations. It's how, uh, you know, troops that aren't co-located with each other communicate. It's how, you know drones, um, you know, are controlled, uh, you know, whether it's. Um, satellite to air or ground to air, uh, communications.
Uh, and, you know, those who have the advantage in the EMS are able to, um, you know, have quote unquote, what we consider military superiority in the battlefield. Um, which is why there's GPS jamming or jamming of all kinds of signals. And to be able to degrade, um, uh, cap military capability. Um, whether it's, uh, our enemy, uh, military capability or our, uh, enemies trying to degrade our capabilities.
So, you know, how do we use Linkurious and how do we use our platform, uh, you know, to be able to bring all this information together and make a, a more effective, um, war fighter to be able to give. The military, um, situational awareness of what's going on in the battlefield. Um, you know, a lot of these systems have been built over the last 20, 30 plus years, so there's a lot of stove pipe data.
There're, uh, disparate data and stove pipe systems. Um, you know, a lot of the correlation of information between those systems, uh, it's done manually. Uh, and then there are a myriad of inconsistent standards. So bringing all of this together is a heavy data engineering and data science lift. Um, you know, we utilize our platform, uh, within the US federal government to be able to bring that information together, uh, in a homogenized form to then be able to provide situational awareness, decision support, and command control. So we're gonna go through a, a, a video of how, um, our system works and our integration, uh, with Linkurious.
EMBM uses a layer approach to displaying entities on the map, allowing users the ability to quickly display, hide, sort, and select EOB objects based on definitions such as affiliation, domain, and type. Users will be able to quickly create EOBs in their workspace based on a OI, and then immediately have access to exploratory and analysis capabilities depending on their needs. Through a combination of layer table and the MAP interactions, users will have the ability to create snapshots of a situation in time. The newly implemented legends window enhances user comprehension. By providing clear intuitive explanations of map elements, users will maintain the ability to view details of any object, including links to associated units, platforms, systems, and equipment. The EMBM Nodal view will also give users a brand new visualization option that highlights relationships between entities such as hierarchy impact and spectrum overlap.
Thank you for that. So you can, as you can see in our system, uh, you know, we're providing, uh, situational awareness, uh, in the electromagnetic spectrum. Um, you know, for the war fighter, um, we're taking, uh, information of, uh, hierarchy, uh, and linkages, uh, between systems, platforms, uh, units, uh, and altogether. So this permits, uh, the military user to understand, uh, given an area of responsibility or an area of interest.
Um, what assets they have in play, what enemy assets might be there. Um, you know, where there are, uh, weaknesses, uh, both in, uh, what we, uh, call high value assets or HVAs, uh, or hvs, are high value targets. Um, being able to look at those linkages and maybe see that, you know, one system is providing a core radar capability, um, or a, a core electronic warfare capability and there are no other backups in the area, um, you know, presents a weakness, uh, if it's, uh, our assets and, you know, presents an opportunity to us if it's an enemy's asset. Um, the technologies that we use.
Uh, that underlie all of this, um, are Linkurious, of course, to be able to perform that visualization analytic. Uh, we, uh, utilize Neo4j, um, you know, as a, um, as a, as a database to be able to maintain that information. Um, we also utilize Databricks as our core platform, uh, for that real time streaming information of military intelligence. Um, you know, the great part of the integration of these components are that. Permits of flexible schema. So as new standards come, uh, you know, are are brought online, you know, we're able to quickly, uh, incorporate that information.
Um, it also allows us to, uh, utilize a domain specific ontology so we're able to convert, uh, and maintain information from these disparate data sources. Uh, in one homogenous form, uh, which makes it easy for tools like Linkurious and our system vore to be able to man manage and maintain, uh, that information. Um, that graph provides a, uh, extensible framework so we're able to quickly add new domains. So our core area of, uh, use for the system is in the spectrum and, uh, on the intelligence side. Something called signal intelligence. Um, but as we incorporate other, uh, data sets like, uh, human intelligence, humans, uh, image intelligence, image, uh, and others, uh, you know, this provides, uh, an, uh, extensible framework for us to quickly be able to correlate, uh, and, uh, you know, build, uh, relationships with those new data sets.
Um, we run this, um, all in the DODs. Uh, cloud network, um, uh, at the secret level. So these, this is on an air gap system. Uh, and, you know, the technologies permit that type of use.
Um, you know, Linkurious, uh, scales as we have additional users, uh, on the DOD Secret Network, um, of the system, and then, um, you know, with partners like, uh, Neo4j and Databricks. Uh, we're also able to, um, horizontally scale, uh, data as, um, you know, we have additional real near realtime sources come into the system. Um, there's sustained performance, uh, so in a traditional SQL or no SQL database, you know, these relationships of, uh, graphs, um, you know, the, um, to be able to traverse. The number of edges that exist between, uh, one node and another, it just wouldn't be possible, uh, in, you know, a traditional, uh, database, uh, or, uh, you know, within, uh, a NoSQL database. So, um, you know, having both, uh, you know, graph database on the backend and, uh, a graph, uh, visualization analytics tool like Linkurious allows us to be able to, uh, display and maintain this complex relationship of information. You know, um, you know, the, one of our core systems that we, uh, utilize, I'm curious for, um, is to provide situational awareness to, um, you know, military users worldwide.
Um, so graphs allow that enhanced awareness, um, you know, integrating, uh, the information that we receive from these disparate sources, uh, into Neo4j and visualizing it. With Linkurious allows us to be able to communicate the relationships of, of the various pieces of information to the military user base. Uh, whereas, uh, typically they would see a flat relationship, uh, between entities. So if you have, um, a unit of force, um, you know, let's say like a battalion somewhere deployed in the Middle East, uh, and then, uh, you know, you have, um, its equipment, its use of spectrum, uh, et cetera, that's, uh, that's. You know, quite complicated to be able to visualize, uh, easily to users. But, uh, utilizing Linkurious, we're able to show that information as a graph and then allow, uh, the users to easily traverse, uh, the varied relationships.
Um, you know, going to clarity, speed, and insights by seeing the varied relationships. Um, this allows, uh, users to be able to understand, uh, you know, either interoperability or weaknesses. Uh, in, uh, deployed systems for us, and then to also see those things for our enemies.
Um, you know, the, the platform is proven. Um, we also partner with Senzing, who's a Linkurious, um, partner. Uh, they provide the, um, energy resolution, uh, uh, analytics to be able to take multiple entities. You know, again, you're, you know, bringing data in from. Um, you know, two, three, you know, dozens of stove pipe systems and a lot of those systems contain the same information but just slightly different.
Um, you know, by, um, you know, using the integrated, uh, any resolution that Linkurious has with sensing, this allows us to be able to convolve that information and give the users a much cleaner, uh, view of. How varied data is linked and be able to onboard new data sources and use the entity resolution to be able to fuse that information together. Um, and lastly, you know, uh, it, it gives us an operational edge, um, by having that graph based situational awareness. You know, we're able to take this raw information and show, you know, real time relationships, bringing in the links together.
Uh, and then, uh, you know, helping, you know, DOD understand, um, the interrelationship, uh, between systems and other kinds of information. Uh, you know, I thank you for attending this seminar. I think, uh, we'll be going into the Q&A. Momentarily. Um, but really appreciate everyone's time.
All right. Thank you so much. Uh, so indeed we have some time for q and a. Uh, if you have some questions
you haven't popped into the q and a panel, uh, feel free to drop those in. Uh, but we can already get started. Uh, so we have a first question from Garrick. Uh, how does the system handle unidentified or ambiguous data elements? Um, so you know, that's, uh, more on our d data engineering side. So we have, uh, two ways of being able to handle that.
Uh, we have a domain specific ontology. So in this space it would be, it's a spectrum operations domain, specific ontology. So as we get information in, uh, we map that information into the ontology, and then from the ontology, now that we have a normalization of.
Uh, data elements of terms and everything else. Uh, you know, we're able to, uh, utilize both the entity resolution, uh, component that we have and some other AI ml that we have. Uh, you know, generative ai. System to be able to take, let's say system parameters and there's no, uh, you know, nomenclature or name of the system. But to look at the parameters and be able to state that, that, that, you know, given the parameters, it's most likely x, y, Z system given where it's operating.
Uh, and you know, the parameter set. Um, you know, we have these type of, um, you know. You know, complex workflows on our data engineering side to help, uh, handle those, uh, unidentified, uh, data pieces. Now, if the, you know, data elements don't, uh, you know, uh, prescribe to the ontology that we have, uh, then we have, uh, a hu you know, human teamed, um, you know, ETLs, where, um, that information will end up in a queue, uh, where our analysts will be able to then, you know, further research, um, you know. The nature of that information.
All right. Thank you so much. Um, so we have another question here that I think is also for you right here. Uh, what was the experience like for the expression team to customize the Linkurious platform for your use case? Um, they, they were actually, um, we were assigned a sales engineer.
Uh, they did, they did a great job. Uh, as I mentioned, you know, our, uh, our core program with them is in on an air gap. Uh. A network. So it's deployed in a classified, uh, DOD network.
So from, you know, the information assurance pieces, uh, to, you know, the configurations to be able to deploy and, uh, manage on that air gap network, you know, where you run this in a Kubernetes cluster. Uh, the Linkurious, uh, technical support team, you know, has, has worked with us hand in hand with our DevOps team, uh, with our development team, uh, and with our cyber team. Um, so the, you know, I, I know, um, you know, we've talked to them about, we're so impressed we've talked to them about a long-term partnership, uh, in, in the federal space. Um, because, you know, they've given us, uh, white glove support end to end. Fantastic.
Thank you. Um, so we have another question here from Kenneth. Um, how does the collaboration between, uh, Linkurious and Expression Networks specifically address the evolving requirements of, uh, JEMS, uh, oh O of gem? So, yeah, gem, sorry.
No, of course. Uh, I mean the evolving requirements in gem, so are, you know, originally when the concept. Came out in 2016, so almost a decade ago. Um, you know, we're looking at a world where, uh, the goal was to bring in the three core disciplines, uh, um, that are used in the military. So spectrum management, which is the licensing and, and kind of maintenance of, uh, friendly communications, right? Like where can I communicate, where can I not let me coordinate licenses so that I know that I have this much spectrum to use in this place, you know, for this time period. Um, the intelligence community that provides.
Collection support, um, you know, from our three letter agencies, uh, and as part of the US government, uh, and then the electronic war, uh, warfare community where, uh, you know, we use, um, you know, electrons to be able to disrupt enemy communication. So it's, you know, taking those three disciplines, uh, in play. Um, and, you know, the, the concept was, you know, came out, uh, in the early, uh, 2000 teens, um, to today, understanding that there's more and more, uh, supportive data. That's required beyond the information from these three domains to be able to accurately understand the maneuver, uh, and the, um, how to exploit, uh, the electromagnetic spectrum. So by u utilizing Linkurious, and, uh, you know, graph technology on the back end. We're able to, you know, quickly and rapidly bring on these additional data sources and these data domains to be able to support the understanding of where the weaknesses lie.
Not just in our standard understanding of, you know, uh, communication paths, but also, you know, the, the different support elements, right? So if I can't, um, communicate for, um, you know, between A and b, then maybe my logistics supply chain goes down. If I can't, uh, listen to, you know, uh, node C, which is some enemy node from an intelligence community standpoint, then I lose this type of awareness of enemy plans. So, um, you know, it's, it's become, it's helped us, helped the community expand their understanding and their ability to add on these varied support domains. You know, quickly, rather than, you know, building a whole new system to be able to, you know, take the human intelligence, uh, view or the image intelligence view, now we can just bring on that data, you know, create the appropriate edges, uh, within the graph, and then show the interrelationships between that information.
Super interesting. Thank you. Uh. A little bit similar perhaps, but uh, really related to the visualization specifically. Um, so the question is, uh, can link curious platform be customized to display specific, uh, JEMSO related data visualizations and metrics that are critical for particular military operations or training exercises? A hundred percent. So, you know, we've been expanding it so that, um, you know, we can show the, the very, you know.
Um, mill Standard 25, 25 iconography for, uh, any of the, the objects that, uh, we've, you know, generate within Neo4j and visualize within Linkurious we're able to then show that, you know, um, you know, where they are geospatially. So, you know, we can show a unit geospatially, you know, uh, located, let's say in the Eastern Mediterranean. Uh, and then you can double click on that and then see all the platforms that are associated with that unit, all the equipment that's associated with those platforms. Be able to quickly search within band ranges to understand, um, you know, where you might be vulnerable from, uh, you know, somewhere, uh, like in, in your used, uh, or prescribed frequency range. Um, should there be jammers in that area? Um, you can really quickly see what jammers, uh, can then affect.
Um, those equipment that run in that frequency range, uh, and these are all, you know, first class properties of the link curious platform. So it's, uh, greatly enhanced and enabled us to be able to expand, uh, that situational awareness. Um, like and just within the system, the APIs very extensible.
So, you know, we've been, um, slowly but surely adding these capabilities on their platform. Thank you so much. Uh, all right, we have one last question. Uh, I think it's a nice one to sum it up. Um, so, uh, why did you, uh, in parentheses the expression team pick the Linkurious Platform to tackle this project? So, you know, we, we looked at, um, you know, different platforms.
We looked at Star Dog. I, you know, we used Databricks, you know, as the data back plane for the system. So we looked at Star Dog, we looked at Linkurious, uh.
Uh, we looked at Cambridge, uh, intelligence, um, uh, which, uh, provides a, um, a graph visualization platform. And, Linkurious was just turnkey. It was just out of the box, turnkey.
Uh, it supported our, uh, our data back playing, uh, and, and working with the team, the sales engineering team. They provided so much support in how to integrate and how to, um, you know, merge the curious. Uh, platform to our own to give, um, give a, a seamless, um, ex uh, uh, CX and UX to our user base. Uh, it, it, you know, it became a no brainer for us to integrate when curious over those other platforms. Alright, thank you so much. Uh, so I promised that was the last question, but there's actually another one that came in, so lemme read that one out.
Also, um, can you visualize the interactions between opposing systems at a JS level? Uh, at, uh, at a joint staff level, uh, sorry. Uh, the Js JJ J slash O, um, yeah, no, uh, we can, so, uh, what we do is, um, you know, we, we actually partner with a company called STK, which is owned by Ansys, um, to do the, uh, engineering, uh, portion of that. Um, and so we're in the one Curious platform. Uh, we're able to make the integration of, uh, that engineering service to be able to build edges, uh, and, and nodes, um, on, you know, whether, um, you know, uh, you know, like an S to N or a J to S. Um, you know, uh, thresholds met are not met and dynamically, you know, build an edge, you know, stating that there, there'll be degraded communications or there'll be, uh, uh, clear communications and then visualize all of that, uh, within when curious with, you know, varied colored edges on whether you have good comms, poor comms, whether, you know, you should, uh, be concerned.
And then that all wraps into their query engine to be able to then, you know, show that in an area. There are, you know, you know, degraded comms, you know, uh, you know, given the engineering service, building those edges and then being visualized within when curious. All right.
Thank you so much. Uh, so I think that's all the questions in the q and a. Um, thank you so much, Abir, for joining us. Uh, it was really great to have your presentation today.
And thanks for all the attendees for being here. Uh, we'll see you at the next webinar. Thank you all. Thanks.
2025-05-02 08:11
