BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20240626T180033Z
LOCATION:Level 2 Lobby
DTSTART;TZID=America/Los_Angeles:20240626T180000
DTEND;TZID=America/Los_Angeles:20240626T190000
UID:dac_DAC 2024_sess256_LBR110@linklings.com
SUMMARY:Late Breaking Results: Circuit-Algorithm Co-design for Learnable A
 udio Analog Front-End
DESCRIPTION:Late Breaking Results Poster\n\nJinhai Hu and Zhongyi Zhang (N
 anyang Technological University), Cong Sheng Leow (University of Michigan)
 , Wang Ling Goh (Nanyang Technological University), and Yuan Gao (Institut
 e of Microelectronics)\n\nThis paper presents a circuit-algorithm co-desig
 n framework for learnable audio analog front-end (AFE) which includes an a
 nalog filterbank for feature extraction and a classifier based on Depthwis
 e Separable Convolutional Neural Network (DSCNN). Instead of the tradition
 al approach to design the analog filterbank and digital classifier separat
 ely, a learnable filterbank is proposed and its source-follower bandpass f
 ilter (SF-BPF) parameters are optimized together with the neural network c
 lassifier in a signal-to-noise ratio (SNR)-aware training process. A new s
 ystem criterion function (LBPF) is proposed to include classification loss
  and filter performance into the training process. The optimized audio AFE
  achieves 10.6% and 11.7% reduction in BPF power and chip area, respective
 ly. Meanwhile, this approach achieved 88.6%–94.5% accuracy for 10-keyword 
 classification task across a wide range of input signal SNR from 5dB to 20
 dB, with only 16k trainable parameters.
END:VEVENT
END:VCALENDAR