Εκτύπωση

Εκδηλώσεις

Εκδήλωση 

Τίτλος:
How to reduce hubness in high dimensional data spaces
Πότε:
04.06.2013 14.00 h - 15.00 h
Πού:
Αίθουσα Εκδηλώσεων, Κεντρικό Κτήριο - Μαρούσι
Κατηγορία:
Παρουσιάσεις - Ομιλίες

Περιγραφή

How to reduce hubness in high dimensional data spaces

Arthur Flexer

Summary: Due to a general problem of measuring distances in high dimensional data spaces, hub objects emerge which tend to be among the k nearest neighbours of a large number of data items. This is a novel aspect of the 'curse of dimensionality' which adversely affects classification and identification performance. We discuss the phenomenon in the context of audio based music recommendation, speaker verification and general machine learning. In particular we show how local and global scaling of distance spaces can decisively improve the hubness situation.

 

Short CV: Arthur Flexer works on enabling computers to listen to music. Therefore his research interests include machine learning, pattern recognition, and intelligent music processing. He is a senior researcher and project manager at the Intelligent Music Processing and Machine Learning Group at the Austrian Research Institute for Artificial Intelligence in Vienna. Flexer has a PhD in psychology from the University of Vienna and is author and co-author of more than fifty scientific publications.


This is a LANGTERRA sponsored talk!
langterra_namedlogo 

Συνοδευτικά αρχεία
Download this file (flexer_langterra_04062013.pdf)flexer_langterra_04062013.pdf[ ]1683 Kb2428 Downloads

Τόπος διεξαγωγής

Χώρος:
Αίθουσα Εκδηλώσεων, Κεντρικό Κτήριο
Οδός:
Αρτέμιδος 6 & Επιδαύρου
ΤΚ:
15125
Πόλη:
Μαρούσι
Χώρα:
Χώρα: gr

Περιγραφή

Χωρίς Περιγραφή