Emi López Abad

At UNED, I work as a Statistics tutor. My background includes a degree in Mathematics, a master's in AI Research, and ongoing studies in Intelligent Systems.

For a Chinese character recognition project submitted to EACL 2027, I developed a reproducible Python and PyTorch pipeline covering 812,872 instances and 3,650 classes. The work used mixed-precision training, multi-seed validation, and preprocessing optimization, improving Macro-F1 by up to 27.4% across the evaluated architectures.

For research presented as a poster at IEEE ISBI 2026, I designed and implemented a cascaded architecture using ViT and Swin Transformer to classify breast cancer histopathology into four classes, reaching approximately 93% balanced accuracy.

I also developed a mathematical model for analyzing uncertainty propagation, error accumulation, and reliability in sequential classification systems; the peer-reviewed work was accepted for presentation at ICAIM 2025. I'm interested in applying my AI and data analysis skills to technology and digital transformation projects.