TY - GEN
T1 - Learning invariances with stationary subspace analysis
AU - Meinecke, Frank C.
AU - Von Bünau, Paul
AU - Kawanabe, Motoaki
AU - Müller, Klaus R.
PY - 2009
Y1 - 2009
N2 - Recently, a novel subspace decomposition method, termed 'Stationary Subspace Analysis' (SSA), has been proposed by Bünau et al. [10]. SSA aims to find a linear projection to a lower dimensional subspace such that the distribution of the projected data does not change over successive epochs or sub-datasets. We show that by modifying the loss function and the optimization procedure we can obtain an algorithm that is both faster and more accurate. We discuss the problem of indeterminacies and provide a lower bound on the number of epochs that is needed. Finally, we show in an experiment with simulated image patches, that SSA can be used favourably in invariance learning.
AB - Recently, a novel subspace decomposition method, termed 'Stationary Subspace Analysis' (SSA), has been proposed by Bünau et al. [10]. SSA aims to find a linear projection to a lower dimensional subspace such that the distribution of the projected data does not change over successive epochs or sub-datasets. We show that by modifying the loss function and the optimization procedure we can obtain an algorithm that is both faster and more accurate. We discuss the problem of indeterminacies and provide a lower bound on the number of epochs that is needed. Finally, we show in an experiment with simulated image patches, that SSA can be used favourably in invariance learning.
UR - http://www.scopus.com/inward/record.url?scp=77953188573&partnerID=8YFLogxK
U2 - 10.1109/ICCVW.2009.5457715
DO - 10.1109/ICCVW.2009.5457715
M3 - Conference contribution
AN - SCOPUS:77953188573
SN - 9781424444427
T3 - 2009 IEEE 12th International Conference on Computer Vision Workshops, ICCV Workshops 2009
SP - 87
EP - 92
BT - 2009 IEEE 12th International Conference on Computer Vision Workshops, ICCV Workshops 2009
T2 - 2009 IEEE 12th International Conference on Computer Vision Workshops, ICCV Workshops 2009
Y2 - 27 September 2009 through 4 October 2009
ER -