Hi there, I have implemented the alternative regulation program learning proposed on our last meeting, as well as a probabilistic assignment of regulators to internal nodes. Initial results look quite interesting. Should we have another meeting to keep you up to date? Tomorrow afternoon or next week monday or tuesday would suit me fine. If you want to have a look at some figures: /nas/biocomp/projects/segal/java/tomic/testresults/test13/modules13.pdf The numbers on each internal node are the score gained by the split and the log probability of the regulator assignment. A yellow regulator means it belongs to the module itself (no cyclicity tests for the moment). Cheers, Tom