In this section you may find the list of software that the MUDI Laboratory has worked / is currently working on.
Human-AI
Interaction
Assessment Tool
Online tool for the assessment of Human-AI Interaction in classification
decision tasks via benefit and reliance diagrams and scores.
Decision Support Quality Assessment Tool
Online calculator for the assessment of the quality of AI-based decision support systems, along different and complementary dimensions, such as robustness, calibration, utility and impact on human decision making.
Weakly Supervised Machine Learning tools
A package featuring utilities and algorithms for weakly supervised ML. Should be (more-or-less) compatible with scikit-learn! It collects original algorithms and methods developed by the contributors, as well as some algorithms available in the literature.
Calibration Framework (Calframe)
Calibration is a multidimensional concept essential for assessing machine learning models. It helps understand a model’s global calibration performance, identify miscalibrated regions of the probability space, and determine the level of overconfidence or underconfidence of a model.
Metrics for machine learning evaluation
This a repository collecting all metrics, algorithms and pieces of code related to data quality for Machine Learning, developed by me and others at the MUDI lab of the DISCo dept. @ University of Milano-Bicocca
