👩🏽💻 PhD Candidate in ML applied to Computational Biology | 🎓Graduated in Physics
🌐 Check out my webpage for more details about my work!
- Benchmark of inductive and transductive node embedding models for drug–target interaction prediction.
- Generalization studies that revealed data leakage issues of transductive models that employed node2vec.
- Design of a novel (biologically-drive) negative subsampling technique.
- This work was published at Nature Machine Intelligence ’25. And Oral at NeurIPS ’23 Workshop.
- Ultrafast GNN-based method.
- Better performance and several orders of magnitude faster than state-of-the-art models.
- Strong generalization across different dataset sizes.
- Preserves biological information in node embeddings.
- This work was published in Bioinformatics '24.
- Programming Languages: Python, R, bash, latex.
- ML/DL Frameworks: PyTorch, PyTorch Geometric, scikit-learn
- Chemical informatics: RDKit, ChEMBL, PubChem API
- Others: Git, Docker, HPC.