I'm interested in supervised methods, including deep learning, to teach machines how to classify (mainly clinical) text.
I study syntactic and semantic properties of clinical texts that affect their understanding by machines.
I'm fascinated by the different aspects that influence ranking in a biomedical search engine and affect precision and recall.
Text Classification
Information Retrieval
Acronym Resolution
ICD Classifier
Search Engine
Grammar Checker
During my bachelor at the Paraná Catholic University (Curitiba, Brazil), I got my first contact with machine learning research and employed active learning to train natural language models for medical discharge summaries.
In 2010, I moved to a larger metropolitan area and did my master's at the top-100 (Times Higher Education) University of São Paulo, with a focus on statistical text classification techniques applied to oncology pathology reports.
In 2011, I decided to go deeper into my data and got a job at the A.C.Camargo Cancer Center (São Paulo, Brazil), where I was also involved with the design and development of a search engine to retrieve eligible patients for clinical trials.
Four years later, I decided to expand horizons and went abroad to do my PhD at the Medical University of Graz (Austria), where I explored deep learning models for information retrieval, text cleansing and text classification in the clinical domain.
michel@oleynik.dev
