Machine Learning Seminar: Leong -- Quantisation of Deep Neural Network Training

The details about the machine learning seminar are as follows: 

Time: Mon 24 Aug (2:00 - 3:00pm): 

Location: SMRI Seminar Room (A12-03-301) A12 Macleay Building, Level 3, Room 301.  

Speaker: Philip Leong (USYD) 

Title: Quantisation of Deep Neural Network Training 

Abstract: As the need to improve energy efficiency for AI applications becomes
increasingly important, hardware vendors have reduced the precision of the underlying
computer arithmetic.  In this talk, we discuss quantisation in deep learning and present
block minifloat (BM), a parameterised minifloat format optimised for low-precision,
edge-training applications.  While standard floating-point representations have two
degrees of freedom, the exponent and mantissa, BM exposes an additional exponent bias,
allowing the range of a block to be controlled.  Similar accuracy to floating point can
be achieved with 4-8 bit wordlengths.  

Biography: Philip Leong received the B.Sc., B.E.  and Ph.D.  degrees from the University
of Sydney.  From 1997-2009 he was with the Chinese University of Hong Kong.  He is
currently Professor of Computer Systems in the School of Electrical and Computer
Engineering at the University of Sydney, Chief Technology Officer at CruxML and Chief
Scientist at TernaryNet.