Taiga Abe

My primary research interest is the development of reproducible and robust data analysis tools, with a specific focus on problems in neuroscience. My graduate research has focused on two major projects along these themes:

A Cloud Computing Platform for Reproducible Data Analysis. I am the lead researcher and developer for Neuroscience Cloud Analysis As A Service (NeuroCAAS), a computing platform that removes the burden of computing infrastructure from neuroscience researchers as they use and develop novel data analysis tools. NeuroCAAS achieves this goal by automatically deploying the computing infrastructure necessary to use any given data analysis tool, providing computational reproducibility by design. Learn more about the project here.

The Benefits and Limitations of Neural Network Ensembles. Ensembling is a popular benchmark method to improve the performance of deep neural networks. My work on these “deep ensembles” studies more recent claims that ensembling also improves the robustness of neural network predictions, and investigates the proposed mechanisms behind their success.

I am currently on the job market for a postdoctoral research position in computational neuroscience. As a postdoc, my goal is to build reproducible, robust data analysis tools that can test scientific hypotheses in real-time experimental settings.

Beyond research, I am an avid rock climber (primarily bouldering) and cellist (primarily chamber music).