Bio: Karina Grosheva
Bringing highly trained image annotation teams for specific computer vision cases. I am a CEO of TaQadam that offers a combination – tech tools and API for data exchange – and humans in the loop to deliver image annotation in the simple and secure way.
We are driven by a social mission to create an infrastructure for the ‘work of the future’ in the Middle East and North Africa. We believe that machines won’t take jobs they can also help train us for new ones. No profession will be exempt from working with AI. We are working on tools and systems to bring such ‘new work’ to youth in the region.
Specialties: Artificial Intelligence, Computer Vision, Social innovation, United Nations, International Development, Humanitarian work, Refugees. I read moral philosophy and rise for global economic inclusion and dignity for all.
TaQadam works in the space of satellite and drones imagery analysis, we develop ai for image process optimization (classification and object detection), annotation tools/mapping tools. The process also involves human-in-the-loop, which includes both data curation/mapping at the same time. Specifically trained and collaborative teams that focus on the developing of consensus in often hard to annotate/indistinguishable objects in sats. We have done a project on drones imagery in disaster response, and agrotech here in the US annotating and pre-processing more than 5 million images.
Title of Talk: Responding to complex social and business challenges with Earth observation data and computer vision.
Geospatial AI has been used across business industries in developed countries, including agriculture, insurance and asset monitoring, energy sector. It is also becoming a major source of data in fragile countries of insufficient statistical capacity, often to solve complex challenges, such as mapping critical infrastructure or responding to humanitarian challenges. While national spatial agencies often lack high resolutions satellite data, and commercial data is expensive, a combination of tools, human in the loop annotation services and computer vision, allow make these projects possible.