Important: Note that this seminar is on a Thursday 10.00-11.00
Title
Uncertainty quantification through conformal regressors and predictive systems
Abstract
Conformal prediction is a framework that allows for controlling the error rate of any predictive model. This is achieved by turning point predictions into set predictions, which are guaranteed to include the true target with a probability that is specified by the user. This guarantee is provided without any model or distributional assumptions, except for that of exchangeability, which is weaker than the standard IID assumption. In this talk, we will focus on conformal regressors, which transform point predictions of any underlying regression model into prediction intervals, and on a recent generalization of these, called conformal predictive systems, which produce cumulative distribution functions. From the latter, prediction intervals can be obtained, as well as percentiles, calibrated point predictions, and p-values for given target values. Standard, normalized and Mondrian conformal regressors and predictive systems will be covered in the talk. The techniques will be illustrated using the crepes Python package.
Bio
Henrik Boström is Professor of computer science - data science systems, KTH Royal Institute of Technology, Sweden. His research concerns trustworthy machine learning, and focuses in particular on predicting with confidence (conformal prediction) and explainable machine learning. As results from — and tools for — this research, he has developed the Python packages crepes (Conformal classifiers, regressors and predictive systems) and xrf (Explainable random forests).
Practical
Speaker is presenting in person.
This guest talk is part of the seminar series on the thematic focus of uncertainty quantification (Objective 4: Machine learning with quantified uncertainty).
Open for attendance.
For those unable to attend in person: https://uio.zoom.us/j/62535744013
Point of contacts:
- Organiser and moderator: Professor Ingrid K. Glad, University of Oslo
- Integreat Administrative Leader Maria Dikova, University of Oslo