Victor Coscrato

Victor Coscrato is a Master in statistics by UFSCar/USP and Bachelor in Statistics byUFSCar. He was born in São Paulo. During his BSc he conducted his first scientific research on agnostic statistical inference. During this time, he was a volunteer member in both the tutorial education program in statistics (PET-EST/UFSCar) and the Risk study centre (CER/UFSCar). His BSc graduation work consisted of a comparative study between word2vec and bag-of-words on natural language processing classification applications. In the last period of his BSc, Victor worked as an intern in people analytics for Stone Company in São Paulo. His job was to create automatic mechanisms to accelerate and improve quality of the company’s employee hiring process. During his MSc jorney Victor discovered his passion in machine learning, which became his main research field since then, leading him to join the statistical machine learning lab (SMaLL/UFSCar). During this time, he wrote two method proposal papers on neural networks usage on regression problems: the first method consisting in a neural network regression ensemble and the latter in a novel interface to local linear smoothing. The MSc dissertation was composed by these two papers combined in a fashion to expose how they connect to each other though. Currently, Victor is interested in general machine learning and artificial intelligence problems, more specifically, he is interested in recommender systems and machine learning models interpretations / prediction explanations. Nevertheless, non-parametric and agnostic statistical inference are still attractive fields for him.

Supervisors: Dr Derek Bridge