Bio

I'm André Gonçalves, a machine learning research scientist. I obtained my Ph.D. in Computer Engineering from the University of Campinas (Unicamp), Brazil in 2016. My thesis was elected as the best Ph.D. thesis in Computer Engineering from Unicamp in that year. I spent half of my Ph.D. at the University of Minnesota, where I worked with Prof. Arindam Banerjee on new machine learning models for a variety of applications including image classification, Alzheimer's disease progression assessment, and climate forecast. It resulted in high-quality peer-reviewed scientific publications. My thesis topic is multi-task learning, a machine learning paradigm in which multiple models are jointly trained in a share-inducing manner so that possible commonalities are leveraged. Currently, I'm designing new and applying state-of-the-art machine/deep learning methods for scientific applications at the Lawrence Livermore National Laboratory. Aside from providing accurate predictions, in such applications, machine learning models must also be capable of producing insights and possible explanations for the scientific phenomena being studied. Produce uncertainty quantification on the predictions is also essential. These requirements pose new challenges to machine learning. I have been privileged to work on exciting projects throughout my professional journey, where I was instigated to learn and keep updated with many sub-fields in machine learning including, predictive modeling, time series prediction, image classification, signal (audio) processing, recommender systems, natural language processing, generative models, and representation learning. Specialties: Machine learning, Deep Learning, Data Mining, Statistical Modeling, Pattern Recognition, Audio (voice) Processing, Optimization, Bayesian Statistical Modeling. Tools: Python (including TensorFlow and Pytorch), Matlab, C++, R.