Russian authorities continue to conceal the scale of their military losses in the full-scale invasion of Ukraine. IStories found a way to reveal it: with the help of AI, they made an algorithm which helps to count it, combine in the database and update it regularly.
The algorithm is trained to collect all the public messages about Russian soldiers killed or missing in the war in Ukraine and extract all the important info from the post: region, date of birth and death, place of death and other info. Is it possible to make this process less painful and long? Come and hear some of the problems the team had to tackle:
- At first, AI thought that actor Alan Rickman had died in the war in Ukraine. How to define the actual war casualties?
- We detected more than 105K Russian soldiers died or missing on the war in Ukraine. The data of this scale can easily be corrupted. How to detect fakes among all of these messages?
- Time, resources, people: How to manage this kind of projects?
- How this experience can be useful for others, who want to reveal something hidden in the depth of social media or other data.
The team will share the database with the session participants and demonstrate how to use it.
