helsinki1.jpg

We have also collaborated with people doing experimental vision research. typically on topics closely related to our computational expertise.

One topic is classification images, which is a method in visual psychophysics to estimate the visual template used by human observers using linear regression, not unlike in reverse correlation. We have also been involved in brain imaging experiments related to vision.

See this page for publications.

Aapo Hyvärinen and Johan Himberg

We collaborate with people working on brain imaging to develop exploratory (unsupervides) data analysis methods for functional brain imaging data.

One often needs to estimate statistical models where the probability density function is known only up to a multiplicative normalization constant. While we encounter this problem in our statistical models of visual processing, it is, in fact, a very general problem in statistical estimation.

Several authors (Spirtes et al. 2000; Pearl 2000) have recently formalized concepts related to causality using probability distributions defined on directed acyclic graphs. This line of research emphasizes the importance of understanding the process which generated the data, rather than only characterizing the joint distribution of the observed variables.