TY - GEN
T1 - Mini-batch spectral clustering
AU - Han, Yufei
AU - Filippone, Maurizio
N1 - Publisher Copyright:
© 2017 IEEE.
PY - 2017/6/30
Y1 - 2017/6/30
N2 - The cost of computing the spectrum of Laplacian matrices hinders the application of spectral clustering to large data sets. While approximations recover computational tractability, they can potentially affect clustering performance. This paper proposes a practical approach to learn spectral clustering, where the spectrum of the Laplacian is recovered following a constrained optimization problem that we solve using adaptive mini-batch-based stochastic gradient optimization on Stiefel manifolds. Crucially, the proposed approach is formulated so that the memory footprint of the algorithm is low, the cost of each iteration is linear in the number of samples, and convergence to critical points of the objective function is guaranteed. Extensive experimental validation on data sets with up to half a million samples demonstrate its scalability and its ability to outperform state-of-the-art approximate methods to learn spectral clustering for a given computational budget.
AB - The cost of computing the spectrum of Laplacian matrices hinders the application of spectral clustering to large data sets. While approximations recover computational tractability, they can potentially affect clustering performance. This paper proposes a practical approach to learn spectral clustering, where the spectrum of the Laplacian is recovered following a constrained optimization problem that we solve using adaptive mini-batch-based stochastic gradient optimization on Stiefel manifolds. Crucially, the proposed approach is formulated so that the memory footprint of the algorithm is low, the cost of each iteration is linear in the number of samples, and convergence to critical points of the objective function is guaranteed. Extensive experimental validation on data sets with up to half a million samples demonstrate its scalability and its ability to outperform state-of-the-art approximate methods to learn spectral clustering for a given computational budget.
UR - http://www.scopus.com/inward/record.url?scp=85031002933&partnerID=8YFLogxK
U2 - 10.1109/IJCNN.2017.7966346
DO - 10.1109/IJCNN.2017.7966346
M3 - Conference contribution
AN - SCOPUS:85031002933
T3 - Proceedings of the International Joint Conference on Neural Networks
SP - 3888
EP - 3895
BT - 2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2017 International Joint Conference on Neural Networks, IJCNN 2017
Y2 - 14 May 2017 through 19 May 2017
ER -