On a mission to responsibly build machine learning predictive models.
Responsible and sustainable predictive modelling is still a new and developing area. We are conducting a number of studies in this domain that examine predictive models applied to tabular data, computer vision or natural language processing models. We investigate the stability and robustness of various methods, work on explainability and transparency for simple and complex models.
As part of our this effort, we develop open source software packages (usually in R and Python) for model explanatory analysis, publish scientific articles describing new methods or investigating properties of already known methods, and create educational materials, recommendations and examples of application in specific domains.
If you want to find out more about what we are working on, check out our seminar, which is always open to those interested in responsible and sustainable data science.