TGN10+ | Research at ISI (Information Sciences Institute), USC | [GitHub] [paper 1] [paper 2]
A framework that takes high-level questions, automatically elaborates them, and run the necessary analyses on the health data (behavioral, psychological, and physiological) of the healthcare workers with the end goal of helping improve their work conditions. It's named after my Granny, Tereatha Grant Nwankwọ (Retired Resigerd Nurse (RN)) with the 10 coming from 10 family members on my Granny's side in the medical field and + with more to come.
A library that include basic time series calculations of interest (ie: autocorrelation, stationarity, etc) mixed with functions to translate time seires into machine learning problems along with forecasting models for both time series and machine learning.
CSCI-535: Multimodal Probabilistic Learning of Human Communication | Course at USC | [GitHub]
In this research-based course, we (1) learned how to carefully and critically read CS-based research papers and answer specific questions, (2) presented a published work within the scope of the class, (3) implemented several projects (Fleiss’ Kappa for inter-rater agreement, Emotion classification, Fairness and Bias, and statistical analyses), and (4) completed a group project. Futhermore on (4), we leveraged techniques in machine learning to extract and analyze three main modalities- (1) Audio, (2) Text, and (3) Visual in various settings (ie: solo, dyadic, multiparty).
CSCI-567: Machine Learning | Course at USC | [GitHub]
In this blend of theory and application-based course, we learned fundamental techniques along with implemented them. The techniques were specifically implemented with code are (1) K-nearest neighbor, (2) Regression, (3) Classification, (4) Neural Networks, (5) K-means, and (6) Hidden Markov Models. For those that haven't taken this course, know that it relies heavily on a fundamental mathematics of linear algebra, calculus, probability, statistics, and optimization.