講座題目：Parsimonious Extreme Learning Machine Using
Recursive Orthogonal Least Squares
主講人:Prof. Er Meng Joo,
新加坡南洋理工大學智能系統中心主任
講座時間：2014年6月3日上午10: 00
講座地點：前衛南校區計算機大樓A521報告廳
講座簡介：
Abstract:
In this talk, a generalised single hidden-layer feedforward network (GSLFN) is proposed.The GSLFN is extended from the SLFN by employing polynomial functions ofinputs as output weights connecting randomlygenerated hidden units with the corresponding output nodes. The salient features of the proposed GSLFN are:
(1) A primal GSLFN(P-GSLFN) is implemented by using randomly generated hidden nodes and polynomial output weights whereby the regression matrix is augmeneted by full or partial input variables and only polynomial coefficients are to be estimated. (2) A simplified GSLFN (S-GSLFN) is realized by decomposing the polynomial output weights of the P-GSLFN into randomly generated polynomial nodes and tunable output weights.
(3) Both P-GSLFN and S-GSLFN are able to achieve universal approximation if the output weights are tuned by ridge regression estimators. By virtue of the developed batch and online sequential ridge ELM (BR-ELM and OSR-ELM) learning algorithms, high performance of the proposed GSLFNs in terms of generalisation and learning speed is guaranteed. Comprehensive comparative studies with SLFNs on real-world regression benchmark problems are carried out.
Simulation results demonstrate that the proposed GSLFNs using BR-ELM and OSR-ELM are superior to SLFNS in terms of accuracy, training speed and structure compactness.
