My research interests include Machine Learning, Kernel Machines, Speech and Language Understanding, OCR and Document Image Processing and Compression.

I am currently working on the theory and implementation of Kernel and Margin Classifiers for speech and language applications.

Slides about Kernel and Margin classifiers from the ICASSP'05 tutorial on Machine Learning for Speech and Language Processing .

I designed and implemented the segmenter for the DjVu document image compression system that allows to distribute high-resolution scanned documents in color on the Web.
 

Patrick Haffner    
Speech Algorithms Dept.,  
Voice Enabled Services Research Lab  
AT&T Labs - Research   
Room A54E02, 200 Laurel Ave, Middletown, NJ 07748   
tel: (732) 420 9125 fax: (732) 368 9475  
Room E109, 180 Park Ave, P.O. BOX 971, FLORHAM PARK, NJ 07932   
tel: (973) 360 7018   
haffner AT research DOT att DOT com