UFAZ Graduate Ali Khudiyev Successfully Defended PhD Dissertation

Ali Khudiyev, a graduate of the French-Azerbaijani University (UFAZ), has successfully defended his doctoral research in Artificial Intelligence at the University of Strasbourg (UNISTRA) within the framework of the Cotutelle Program, marking a new milestone in his scientific career. He completed his Bachelor’s degree in Computer Science and Master’s degree in Data Science and Artificial Intelligence at UFAZ. Later, with the support of the French government, he pursued his doctoral studies at UNISTRA.

His PhD dissertation, titled “Scaling Intelligence: A Formal and Practical Framework for Computationally Unbounded AI,” focused on developing autonomous AI systems capable of self-scaling and continuous learning. The research introduces the concept of Computationally Unbounded AI (CUAI). Inspired by biological principles of modularity and plasticity, the dissertation aims to enable AI models to autonomously expand and optimize in dynamic and open-ended environments.

Within this framework, Ali Khudiyev proposed two novel methods:

1. Column Extension Framework – scales artificial neural networks (ANNs) through graph-based expansion; subgraphs within the network are selected and replaced with duplicates of the original network, increasing depth and breadth, thereby enabling autonomous growth. This method was validated on the MNIST-1D dataset.
2. Capacity-Aware Learning (CAL) Framework – combines boosting with selective prediction. First, base learners are trained to reconstruct inputs, with a rejection option added. Then, identical models are progressively added to prioritize more difficult samples. Inspired by cortical columns studied in neuroscience, the CAL framework ensures efficient use of resources and adaptive learning, achieving higher reconstruction accuracy and task coverage compared to single models.

During his defense, Ali Khudiyev emphasized that this research provides a conceptual foundation for future AI models that are both more explainable and capable of self-optimization. According to him, much work remains to be done in this field, and his research will continue.

The research results have been presented at international conferences in France and Japan and published in relevant international journals.

The UFAZ community congratulates Dr. Ali Khudiyev and wishes him continued success in his scientific endeavors.