PASCAL Agnostic Learning
vs.
Prior Knowledge
IJCNN07

The challenge is now over. But it remains open for post-challenge submissions!


IMPORTANT: Entries made since February 1st 2007 might be using validation data, now available for training.

EI

Submitted by Vladimir Nikulin

EI -Experience-Innovation (was motivated by RF and RS)
Also, I can report that during last 2-3 weeks I had made a very
good progress towards the JMLR-paper.

Dataset Balanced Error Area Under Curve  
Train Valid Test Train Valid Test
ada 0.1757 0.1948 0.1916 0.8239 0.8022 0.7968 agnostic
gina 0 0.038 0.0574 1 0.9682 0.9496 agnostic
hiva 0 0.2322 0.3045 1 0.7861 0.741 agnostic
nova 0.0114 0.044 0.0614 0.9901 0.964 0.9367 agnostic
sylva 0.0066 0.0049 0.0087 0.9962 0.9983 0.9966 agnostic
Overall 0.0387 0.1028 0.1247 0.962 0.9038 0.8841 agnostic

This entry is a complete agnostic learning entry.