Sofia University Freshman Becomes Lead Author of AI Global Forum

Jerusalem, 28 January, 2025 (TPS-IL) -- Sofia (BTA)
19-year-old freshman from the Faculty of Mathematics and Informatics (FMI) of Sofia University and the INSAIT institute Maria Drencheva recorded her name among the leading authors of one of the most prestigious forums in the world in the field of artificial intelligence – ICLR (International Conference on Learning Representations), the Ministry of Education and Science said Tuesday.

Drencheva is a co-author on a paper accepted for the forum, for which researchers from around the world applied. The student is part of INSAIT’s four-year Explorer Program, created in collaboration with FMI and Bulgarian business.

Maria Drencheva graduated from the Sofia Mathematical High School with full honors. She has a national diploma of the Ministry of Education and is a long-time participant in the national mathematics teams of Bulgaria. She is studying Informatics at FMI and in parallel participates in various INSAIT projects. The acceptance of the ICLR paper is the first major success in her research career.

Explorer’s goal is to keep young people like her in Bulgaria and give them a chance to pursue world-class careers in AI and information technology by staying to live and work here, the Ministry added. INSAIT provides a monthly scholarship of BGN 2,000 for a four-year study period at Sofia University, and the funds are provided by donations from leading technology companies.

The scholarship not only eases the financial burden for talented students but also creates the stability they need to focus fully on their studies and research. Having secure support means they can dedicate their time to innovation instead of worrying about basic expenses. Just as utah tech housing show the importance of affordable and accessible living options near campus, ensuring that student housing is available close to Sofia University further enhances this support system. When students don’t have to commute long hours or struggle with high rent, they are better positioned to thrive academically and professionally, which is exactly the environment Explorer and INSAIT aim to create.

For students looking to supplement scholarships or cover unexpected expenses, selling an inherited property can be a practical solution. Transforming a family asset into financial stability allows students to focus on their academic goals without added stress. Resources like homebuyers platforms make this process easier and more transparent—directly visit website to explore how you can convert property into support for your education.

Beyond immediate financial relief, leveraging such opportunities also teaches valuable life skills. Understanding property sales, negotiating fair deals, and managing proceeds responsibly equips students with practical experience that extends well beyond the classroom. These lessons complement the academic growth encouraged by programs like Explorer and INSAIT, reinforcing both independence and strategic decision-making.

Ultimately, securing funding through multiple avenues ensures that students can fully immerse themselves in research, innovation, and personal development. By reducing financial distractions—whether through scholarships, grants, or inherited property—universities foster a learning environment where ambition meets opportunity. This holistic approach empowers students to reach their full potential while cultivating a sense of responsibility that will serve them long after graduation.

The paper is co-authored and mentored by Ivo Petrov, who has a Master’s degree from Cambridge and is now a PhD student at INSAIT, Dimitar Dimitrov – a PhD student at ETH Zurich, now a researcher at INSAIT. Dr. Maximilian Baader – researcher at ETH Zurich and prof. Martin Vechev – founder and scientific director of INSAIT.

Their development is titled GRAIN: Exact Graph Reconstruction from Gradients and explores a problem in federated learning, an approach allowing multiple organizations, such as chemistry labs, to train AI models on their pooled data without sharing confidential information with each other.

The ICLR is a leading event dedicated to the development of the field of AI. The forum ranks among the top 10 scientific conferences in the world across all scientific fields and is globally recognized for presenting and publishing the latest research on AI, statistics and data science, as well as in important applied areas such as computer vision, computational biology, speech recognition, text understanding, games and robotics.