Building predictive models that are a hybrid of data driven statistical machine learning and domain knowledge.
Through Hybrid Prediction, we can reimagine existing approaches with new ones to create new opportunities and AI technologies for decision-making in real-time environments.
Hybrid Prediction focuses on including physical or theoretical constraints such as conservation of momentum, mass and energy into data driven models.
Hybrid Prediction can result in replacing components of biophysical models with data driven counterparts, using ML to make them run faster.
Hybrid Prediction is about incorporating everything we know or observe about a problem to generate predictions that are more efficient and more accurate. We use a hybrid of solutions from Statistics and ML to achieve this.
Chat with Petra
- I’m excited to be leading a core team of researchers in CSIRO with specialist skills in statistics, computing and machine learning to solve challenging problems in the Australian landscape. You can connect with me more at https://people.csiro.au/K/P/Petra-Kuhnert.
An Adaptive Solver for Systems of Linear Equations
Computational implementations for solving systems of linear equations often rely on a one-size-fits-all approach based on LU decomposition of dense matrices stored in column-major format.