Short Bio

My name is Gustavo Batista and I am the head of the Machine Learning group at the School of Computer Science and Engineering at the University of New South Wales.

I have researched machine learning for nearly 30 years, which has given me a comprehensive understanding of the area. I have diverse interests. My previous research concerns data pre-processing and evaluation methods, and I investigated and proposed techniques for missing data imputation, class imbalance treatment, and model assessment. I have recently developed methods to learn from time-oriented data, including time series and data streams.

My research in data streams led me to develop methods that identify changes in data distributions and adapt classifiers to those changes. I have developed methods for label shift detection and adaptation and algorithms to estimate the class prevalence, a machine learning task known as quantification.

While I am passionate about theory, I find it equally exciting to translate these concepts into applications. For example, I developed an optical sensor to identify insect species using wing movement. This technology can be applied to monitor disease-vector mosquitoes with funding from USAID (Zika) and IVCC-UK (Malaria). I have also developed object-detection models that are efficient to execute on nanosatellites’ onboard hardware.

I joined UNSW as an associate professor in 2019 after working for more than ten years at the University of Sao Paulo (USP).