Monday, 11th June
Tutorials: Aula, School of Business and Economics, Maastricht University
Kolmogorov lecture: Aula Minderbroedersberg of Maastricht University
Registration
Symposium Opening
Tutorial 1
Confidence Prediction: Introduction to Conformal Prediction, by Lars Carlsson, Stena Line, Sweden & Henrik Linusson, University of Borås, Sweden
Coffee Break
Tutorial 2
Probabilistic Prediction: Venn-ABERS Prediction, by Paolo Toccaceli, Royal Holloway University of London, UK
Lunch
Tutorial 3
Conformal Instance Transfer, by Shuang Zhou, Chengdu University of Information Technology, China
Coffee Break
Introduction to the Kolmogorov Lecture
Prof. Alex Gammerman
Prof. Vladimir Vapnik
"Rethinking Statistical Learning Theory: Learning Using Statistical Invariants"
In the talk, I will considers Teacher-Student interaction in learning processes. I will introduce a new learning paradigm, called Learning Using Statistical Invariants (LUSI), which is different from the classical one.
In the classical paradigm, learning machine constructs, using data, a classification or regression function that minimizes the expected loss; it is thus data-driven learning. In the LUSI paradigm, in order to construct the desired classification or regression function using both data and Teacher's input, learning machine computes statistical invariants that are specific for the problem, and then minimizes the expected loss in a way that preserves these invariants; so it is both data- and intelligence-driven learning.
From a mathematical point of view, methods of the classical paradigm employ mechanisms of strong convergence of approximations to the desired function, whereas methods of the new paradigm employ both strong and weak convergence mechanisms. This can significantly increase the rate of convergence.
Award Ceremony
Reception Grand Café Soiron
(100 meters to the main venue)
City-guided walk
Tuesday, 12th June
Aula, School of Business and Economics, Maastricht University
Plenary Session 1: Predictive Distribution
Chair: Ralf Peeters
"Cross conformal predictive distributions".
Vladimir Vovk, Ilia Nouretdinov, Valery Manokhin and Alexander Gammerman
[Download PDF]
"Conformal predictive decision making".
Vladimir Vovk and Claus Bendtsen
[Download PDF]
"Inductive Venn-Abers Predictive Distribution".
Ilia Nouretdinov, Denis Volkhonskiy, Pitt Lim, Paolo Toccaceli and Alexander Gammerman
[Download PDF]
Coffee Break
Plenary Session 2: Drug Discovery
Chair: Lars Carlsson
"Using Venn-ABERS Predictors to assess Cardio-Vascular Risk".
Ernst Ahlberg, Ruben Buendia and Lars Carlsson
[Download PDF]
"Conformal Prediction in Learning Under Privileged Information Paradigm with Applications in Drug Discovery".
Niharika Gauraha, Lars Carlsson and Ola Spjuth
[Download PDF]
"Venn-Abers Predictors for Improved Compound Iterative Screening in Drug Discovery".
Ruben Buendia, Ola Engkvist, Lars Carlsson, Thierry Kogej and Ernst Ahlberg
[Download PDF]
Lunch
Prof. Peter Grünwald
"Safe Testing"
In recent years, standard p-value based hypothesis testing has come under intense scrutiny. One of its many problems is the following: if our test result is promising but nonconclusive (say, p = 0.07) we cannot simply decide to gather a new batch of data. While this practice is ubiquitous in science, it invalidates p-values and error guarantees.
Here we propose an alternative hypothesis testing methodology that allows us to do so after all. For simple null hypotheses, our proposal coincides with earlier work by Vovk and collaborators, and amounts to using Bayes factors and/or general test martingales. Here, we work out the composite null case, which allows us to formulate safe, nonasymptotic versions of the most popular tests such as the t-test and the chi square tests. Safe tests for composite H0 are not always Bayesian or test martingale-based but rather based on the Barron-Li 'reverse information projection'. A central innovation is to distinguish between 'optional stopping' and 'optional continuation' which may have repercussions in the conformal prediction world as well.
Coffee Break
Plenary Session 3: Posters
Symposium Dinner
(ship on Maas; under negotiation)
Wednesday, 13th June
Aula, School of Business and Economics, Maastricht University
Plenary Session 4: Other Applications of Conformal Prediction
Chair: Harris Papadopoulos
"Detecting seizures in EEG recordings using Conformal Prediction".
Charalambos Eliades and Harris Papadopoulos
[Download PDF]
"Cover Your Cough: Detection of Respiratory Events with Confidence Using a Smartwatch".
Khuong An Nguyen and Zhiyuan Luo
[Download PDF]
"Conformal Stacked Weather Forecasting".
Jelmer Neeven and Evgueni Smirnov
[Download PDF]
Coffee Break
Plenary Session 5: Transfer Learning, Feature Selection
Chair: Evgueni Smirnov
"Conformal Feature-Selection Wrappers for Instance Transfer".
Shuang Zhou, Evgueni Smirnov, Gijs Schoenmakers, Ralf Peeters and Tao Jiang
[Download PDF]
"Exchangeability Martingales for Selecting Features in Anomaly Detection".
Giovanni Cherubin, Adrian Baldwin and Jonathan Griffin
[Download PDF]
"Transfer learning for the Probabilistic Classification Vector Machine".
Christoph Raab and Frank-Michael Schleif
[Download PDF]
Lunch
Plenary Session 6: Conformal Prediction and Other Methods
Chair: Volodya Vovk
"Conformal Prediction in Manifold Learning".
Alexander Kuleshov, Alexander Berstein and Evgeny Burnaev
[Download PDF]
"Venn Predictors for Well-Calibrated Probability Estimation Trees".
Ulf Johansson, Henrik Bostrom, Henrik Linusson, Tuwe Lofstrom, Haken Sundell and Anders Gidenstam
[Download PDF]
"Aggregating Strategies for Long-term Forecasting".
Alexander Korotin, Vladimir V'Yugin and Evgeny Burnaev
[Download PDF]
"Interpolation error of Gaussian process regression for misspecified case".
Alexey Zaytsev, Evgenya Romanenkova and Dmitry Ermilov
[Download PDF]
Coffee Break
Discussions & Closing Ceremony