Evaluate and apply multimodal models using machine/deep learning techniques to multi-sensor tools for complex biological datasets.

We exploit multi-task learning to improve the screening of mental health conditions.

We leverage state-of-the-art transfer learning models for depression screening using as input the recorded audio, text transcript, and images of patient interview.

We use transfer learning models using audio as inputs for depression screening.

We apply deep learning models to impute missing values and classify multivariate time series, using some modify Recurrent Neural Network. The application of these techniques is in Healthcare.

We use the Web Scraping technique to collect data from real estate website. Then, we compute a new index that is useful as a contrast index.

We developed a methodology to estimate the national value of a property separated by construction and land. These results will improve the Chilean household statistics.

Research assistant, Fondecyt 11090005

Research assistant, Fondecyt 1110334