Transductive learning is the problem of designing learning machines that succesfully generalize only on a given set of input patterns. In this paper we begin the study towards the extension of Extreme Learning Machine (ELM) theory to the transductive setting, focusing on the binary classification case. To this end, we analyze previous work on Transductive Support Vector Machines (TSVM) learning, and introduce the Transductive ELM (TELM) model. Contrary to TSVM, we show that the optimization of TELM results in a purely combinatorial search over the unknown labels. Some preliminary results on an artifical dataset show substained improvements with respect to a standard ELM model. © Springer International Publishing Switzerland 2014.
A preliminary study on Transductive Extreme Learning Machines / Scardapane, Simone; Comminiello, Danilo; Scarpiniti, Michele; Uncini, Aurelio. - 26(2014), pp. 25-32. ((Intervento presentato al convegno 23rd Workshop of the Italian Neural Networks Society, WIRN 2013 tenutosi a Vietri sul Mare, Salerno nel 23 May 2013 through 24 May 2013. - SMART INNOVATION, SYSTEMS AND TECHNOLOGIES. [10.1007/978-3-319-04129-2_3].
A preliminary study on Transductive Extreme Learning Machines
SCARDAPANE, SIMONE;COMMINIELLO, DANILO;SCARPINITI, MICHELE;UNCINI, Aurelio
2014
Abstract
Transductive learning is the problem of designing learning machines that succesfully generalize only on a given set of input patterns. In this paper we begin the study towards the extension of Extreme Learning Machine (ELM) theory to the transductive setting, focusing on the binary classification case. To this end, we analyze previous work on Transductive Support Vector Machines (TSVM) learning, and introduce the Transductive ELM (TELM) model. Contrary to TSVM, we show that the optimization of TELM results in a purely combinatorial search over the unknown labels. Some preliminary results on an artifical dataset show substained improvements with respect to a standard ELM model. © Springer International Publishing Switzerland 2014.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
