Sonor defends his MSc with flying colors! ππ₯³
ππ₯³ Big congratulations to Sonor! He has just successfully defended his MSc thesis, Biomedical Text Classification with Pre-trained Language Models, and scored the highest possible grade! ππ―π
π What was the thesis about? π§Ύ
Biomedical science produces mountains of publications every year ππ. But finding the truly relevant papers, especially those about amyloids 𧩠and antibodies π§ͺ, can be slow and exhausting π΅.
Sonor set out to change that! πβ¨ His thesis explored AI-powered text classification to help researchers filter the noise and find the gems π.
π€ How did he do it? βοΈ
π Compared classical ML models (logistic regression, random forest π²) with transformer-based language models (BioMedBERT, SapBERT, BioMed RoBERTa π§¬π€).
π Built a system that uses titles, abstracts, and journal names to decide if a paper is relevant.
π Automated the process by pulling data straight from PubMed π©Ίπ.
π What did he achieve? π
β¨ Reduced manual curation time drastically β±οΈβ¬οΈ.
β¨ Improved the ratio of relevant papers found by over 350% ππ₯.
β¨ Created a tool that works like a supercharged search engine for biomedical literature π‘π.
In simple words: Sonor built a system that helps scientists find the right papers faster β‘ππͺ.
π Way to go, Sonor! ππ
Weβre so proud to celebrate Sonorβs incredible MSc journey ππ₯. He not only earned the top grade π₯ but also showed how AI and ML can revolutionize biomedical research π§¬π€π‘.
