BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Europe/Zurich BEGIN:DAYLIGHT DTSTART:20220327T030000 TZOFFSETFROM:+0100 TZOFFSETTO:+0200 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3 TZNAME:CEST END:DAYLIGHT BEGIN:STANDARD DTSTART:20211031T020000 TZOFFSETFROM:+0200 TZOFFSETTO:+0100 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10 TZNAME:CET END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20211211T140206Z UID:62E0C828-3B08-40B4-9EA3-257381FE035C DTSTART;TZID=Europe/Zurich:20211201T130000 DTEND;TZID=Europe/Zurich:20211201T140000 DESCRIPTION:Dear members\,\n\nWe will be hosting a lecture given by Prof. K eshab K. Parthi\, an IEEE CAS Distinguished Lecturer\, at ETHZ from 13:00- 14:00\, 01.12.2021. Hope to see you there. A second lecture will be given by Prof. Parthi at EPFL. The room and date information will be sent out la ter.\n\nRegards\,\n\nShih-Chii Liu\n\nSpeaker(s): Prof. Keshab K. Parhi\, \n\nAgenda: \nSpeaker: Prof. Keshab Parthi\n\nDate: 01.12.2021\n\nTime: 13 :00-14:00\n\nPlace: ETF E1\, ETH\, Sternwartstrasse 7\, 8092 Zurich\n\nAbs tract: Machine learning and data analytics continue to expand the fourth i ndustrial revolution and affect many aspects of our lives. The talk will e xplore hardware accelerator architectures for deep neural networks (DNNs). I will present a brief review of history of neural networks (OJCAS-2020). I will talk about our recent work on Perm-DNN based on permuted-diagonal interconnections in deep convolutional neural networks and how structured sparsity can reduce energy consumption associated with memory access in th ese systems (MICRO-2018). I will then talk about reducing latency and memo ry access in accelerator architectures for training DNNs by gradient inter leaving using systolic arrays (ISCAS-2020). Then I will present our recent work on LayerPipe\, an approach for training deep neural networks that le ads to simultaneous intra-layer and inter-layer pipelining (ICCAD-2021). T his approach can increase processor utilization efficiency and increase sp eed of training without increasing communication costs.\n\nRoom: E1\, Bldg : ETF \, ETHZ\, Sternwartstrasse 7\, Zurich\, Switzerland\, Switzerland\, 8092\, Virtual: https://events.vtools.ieee.org/m/289904 LOCATION:Room: E1\, Bldg: ETF \, ETHZ\, Sternwartstrasse 7\, Zurich\, Switz erland\, Switzerland\, 8092\, Virtual: https://events.vtools.ieee.org/m/28 9904 ORGANIZER:shih@ini.uzh.ch SEQUENCE:7 SUMMARY:IEEE SWISS CAS Distinguished Lecture / Accelerator Architectures fo r Deep Neural Networks: Inference and Training URL;VALUE=URI:https://events.vtools.ieee.org/m/289904 X-ALT-DESC:Description: <br /><p>Dear members\,</p>\n<p>We will be hosting a lecture given by Prof. Keshab K. Parthi\, an IEEE CAS Distinguished Lect urer\, at ETHZ from 13:00-14:00\, 01.12.2021. Hope to see you there. A sec ond lecture will be given by Prof. Parthi at EPFL. The room and date infor mation will be sent out later.</p>\n<p>Regards\,</p>\n<p>Shih-Chii Liu</p> <br /><br />Agenda: <br /><p>Speaker: Prof. Keshab Parthi</p>\n<p>Date: 01 .12.2021</p>\n<p>Time: 13:00-14:00</p>\n<p>Place: ETF E1\, ETH\, Sternwart strasse 7\, 8092 Zurich</p>\n<p><br />Abstract: Machine learning and data analytics continue to expand the fourth industrial revolution and affect m any aspects of our lives. The talk will explore hardware accelerator archi tectures for deep neural networks (DNNs). I will present a brief review of history of neural networks (OJCAS-2020). I will talk about our recent wor k on Perm-DNN based on permuted-diagonal interconnections in deep convolut ional neural networks and how structured sparsity can reduce energy consum ption associated with memory access in these systems (MICRO-2018). I will then talk about reducing latency and memory access in accelerator architec tures for training DNNs by gradient interleaving using systolic arrays (IS CAS-2020). Then I will present our recent work on LayerPipe\, an approach for training deep neural networks that leads to simultaneous intra-layer a nd inter-layer pipelining&nbsp\; (ICCAD-2021). This approach can increase processor utilization efficiency and increase speed of training without in creasing communication costs.</p> END:VEVENT END:VCALENDAR