We are proud to share that our paper “Mitigating Model Drift in Developing Economies Using Synthetic Data and Outliers” has been accepted to the NeurIPS 2025 Workshop on Generative AI in Finance.
Read the paper here: https://lnkd.in/dcXy5b7F
This is a historic first: First paper of its kind on this critical intersection of AI, finance, and developing economies.
Among the first-ever contributions from Central Asia to NeurIPS — one of the most prestigious venues in the global AI research community.
Why this matters
NeurIPS (Conference on Neural Information Processing Systems) is widely regarded as the world’s leading AI research conference. Acceptance here signals that the work is at the cutting edge, alongside breakthroughs from MIT, Stanford, Google DeepMind, OpenAI, and others.
For a team from Central Asia to be recognized on this stage is not just a scientific milestone — it is a powerful statement that the region is emerging as a serious contributor to global AI research.
About the Paper
Our research addresses one of the most pressing challenges in financial AI: model drift in developing economies. Traditional models often fail due to limited data and rapidly changing dynamics. We propose a data-centric stabilization method using synthetic outliers and new stability metrics, ensuring more reliable and equitable AI systems for finance in emerging markets.
Why this is huge
• Sets a precedent for AI innovation from underrepresented regions.
• Demonstrates how generative AI can directly support financial stability and inclusion.
• Opens the door for more researchers from Central Asia to engage with the global AI community at the highest level.
Authors & Contributors
Azizjon Azimi, Shuhrat Khalilbekov, Ilyas Varshavskiy, Bonu Boboeva, Sergey Shulgin, Dr. Akhlitdin Nizamitdinov and Haitz Sáez de (Ocáriz) Borde.
This is not just our success — it is a regional milestone. We hope it inspires the next wave of AI researchers from developing economies to push boundaries and make their mark on the global stage.