Year 2021 under the MIcroscope
The end of the year is a good time to summarize. So, what has happened in 2021 in the MI2 group?
Early March starts with the publication of the book Explanatory Model Analysis Explore, Explain, and Examine Predictive Models with examples in R and Python by Przemysław Biecek and Tomasz Burzykowski. The book contains a set of methods and tools designed to build better predictive models and monitor their behavior in a changing environment. The book was more than three years in the making, which is about π times longer than the authors expected ;-) The publisher has agreed to make the online version available free of charge at https://ema.drwhy.ai/.
July was a special time this year for our research group, because the internship started during which 7 projects were carried out, involving 12 students. The projects were of course strongly related to the topics of responsible machine learning, AutoML or data visualization. Read more about these bold projects like tidycharts, forester, fairPAN, Aspects and others at our medium blog https://medium.com/responsibleml.
An important point was the useR! 2021 conference during which we conducted an Introduction to Responsible Machine Learning workshop and many interesting presentations on topics related to responsible machine learning. In preparation for the workshop we developed a mixture of comic book and book that is introducing XAI tools for predictive modelling. The English version is available online https://betaandbit.github.io/RML/ and we are currently working on translations of these materials into Polish, Vietnamese and Turkish.
In mid-September, the paper dalex: Responsible Machine Learning with Interactive Explainability and Fairness in Python was published. Finally, Python users can benefit from a consistent library of tools for XAI and fairness model analysis. This was the third approach, but in the end a very cool and useful package was created based on the experience of previous less successful attempts
October 16–17 is when we participated in the Hack4Med CRACoV. The hackathon was dedicated to creating an innovative tool to automate COVID-19 risk prediction, the so-called covid risk calculator which is expected to provide real support to the medical staff of the University Hospital of Krakow. Based on this two-day event, we started a very cool and fruitful cooperation with fantastic physicians from the Jagiellonian University.
November is the month of the ML in PL 2021 conference, where we presented the results of projects carried out during the student internships. We were present both at the student day, poster session and talks.
At the beginning of December, we started working on xLungs grant!
This is a project funded under the NCBiR Infostrateg program, a strategic program for the development of Polish potential in the field of artificial intelligence. Our goal is to develop a system to support the work of doctors-radiologists, thanks to the enrichment of X-ray images and computed tomography with additional information.
Best wishes for 2022!
If you are interested in other posts about explainable, fair, and responsible ML, follow #ResponsibleML on Medium.
