Artificial Intelligence for Monitoring the War in Ukraine
In an attempt to make its state of the art Artificial Intelligence (AI) analysis software available to anyone looking for validated information on the war in Ukraine, Dutch AI specialist OPT/NET BV has been developing a new application: UKROPTOSS.
This new platform is based on OPT/NET’s successful TSAR-AI data analysis solution. OPT/NET analysis software helps organisations analyse large amounts of data in a fraction of the time normally used by computer algorithms. Now they are unleashing this AI power for analysing, validating and documenting the continuously changing situation in Ukraine.
Rapidly deployable AI
It almost sounds like a rapid deployment defence force. In a way, that is also what it is. Traditional AI systems need a pre-defined model of the situation and pre-labelled learning dataset. Instead, the OPT/NET solution builds on a simple pattern interpretation exercise that fits seamlessly into existing workflows straight away. You can read more about this technology here.
The new UKROPTOSS AI application is based on the evolved TSAR-AI technology, a winner of the Copernicus Masters 2018, which was further developed on a number of projects with ESA and EU funding. It even competed for the IBM ‘AI for Good’ XPRIZE, making it into the third round of that competition.
Now they are giving it back to society by contributing to the coverage of the war in Ukraine. This initiative builds on the commercial products that OPT/NET develops for its customers, but is not a commercial product itself. Development and operations are funded from company resources and its supporters. It will be made available for free to professional users. OPT/NET is hoping for support from its supplier and customer network, either with in-kind help or financial contributions, mainly to cover operational (cloud computing) costs, that increase with increasing numbers of users.
Human language
The big advantage of using an AI layer on top of an enormous amount of constantly changing data is that it understands ‘human’ language. Instead of writing complex queries or scrolling endlessly through complex menus, it understands search phrases like ‘show me all shelling in a 100 kilometre radius of city X’.
Speaking ‘normal English’ is an important enabler for any analytics software to be picked up by large amounts of users. You no longer need to ‘learn’ the software, but you can use it straight away. There is no learning curve. This enables citizen journalists and reporters without a background in geo-analytical systems to use it for reports.
What about data validation?
Earlier this week we wrote about the importance of having access to reliable and verified data. With a lot of unverified and often outright false reports about the war being posted on social media all the time, there is a great need for unbiased, independently validated information.
The UKROPTOSS AI platform gets its primary data feed from the well-known liveuamap.com mapping website, which contains independently verified and geo-tagged reports of activities.
The Live Universal Awareness Map (“Liveuamap”) platform started with the Russian invasion into Ukraine in 2014. It has evolved into a leading independent global news and information site dedicated to factual reporting of a variety of important topics including conflicts, human rights issues, protests, terrorism, weapons deployment, health matters, natural disasters, and weather related stories, among others, from a vast array of sources. (source)
About their validation process, the website says: “At Liveuamap we use proprietary software tools, such as AI web crawlers, to find news-worthy stories which will be forwarded to our group of expert analysts for fact checking. Our dedicated editors decide which facts and stories should be displayed on the map to minimize spam.”
UKOPTOSS AI takes this pre-validated data and adds its AI human-language interface for fast information finding. As it runs independently, it is designed to incorporate other data feeds too, like detailed satellite imagery and SAR data. Using unbiased space data will keep the high standard of verification, while greatly enriching the information.
Read about the truth of satellite imagery
Humanitarian support and journalists
The new application is aimed primarily at professional users. This includes humanitarian support organisations and investigative journalists. Basically anyone who benefits from easy and intuitive access to the latest status of the war.
By combining on-the-ground sources with detailed and frequently updated satellite data, the platform could also become a reliable archive for war crimes investigators. Alleged war crimes, like intentional attacks on civilian targets can easily be validated by comparing reports from sources on the ground with before and after satellite images. This can be useful to (dis)prove whether military installations were present at the attack site or not.
OPT/NET Founder and CEO Taras Matselyukh says: “our goals are purely of humanitarian nature for the recording, documentation, systemisation and later analytics of the events in Ukraine, potentially even for prosecution of war crimes, in case we manage to collect evidence of criminal behaviour.”
Read about space data for war crimes investigations
Calling journalists for beta testing
The UKROPTOSS AI platform is looking to go live some time in the next week. It has just entered an extended beta test phase, for which they are looking for beta testers.
To help them test this phase of the product, OPT/NET are looking for international journalists to register as beta testers. You can get a free account by writing to OPT/NET here.
Calling geospatial analysts for further development
Similarly, OPT/NET would like to call out to geospatial analysts and other (space) data analysis specialists to help them further develop the service and integrate new data sources. Please contact the team at this email address for contact.
About OPT/NET BV
OPT/NET BV is a Dutch company that builds comprehensive AI products. It helps its clients deal with ever increasing volumes of data and complexity. The OPT/NET AI Engine has grown into a series of stand-alone platforms with unlimited potential across a variety of critical and data-intensive industries, most notably Earth Observation.
OPT/NET works to support human analysis of big data challenges, not replacing them. The OPT/NET AI Engine provides domain experts in both structured and unstructured data environments the ability to rapidly develop advanced real-time AI solutions that help them do their jobs more effectively, without requiring an advanced degree in data-science. The OPT/NET solutions are Human-driven, AI-assisted.
All info about OPT/NET AI products can be found here.