BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.2.3.1//EN
TZID:Asia/Kolkata
X-WR-TIMEZONE:Asia/Kolkata
BEGIN:VEVENT
UID:105@cds.iisc.ac.in
DTSTART;TZID=Asia/Kolkata:20250214T160000
DTEND;TZID=Asia/Kolkata:20250214T174500
DTSTAMP:20250207T165833Z
URL:https://cds.iisc.ac.in/events/cds-kiac-seminar-cds-102-14th-february-c
 an-we-make-machine-learning-safe-for-safety-critical-systems/
SUMMARY:CDS-KIAC {Seminar}@ CDS: #102: 14th February: "Can We Make Machine 
 Learning Safe for Safety-Critical Systems?"
DESCRIPTION:We welcome you to CDS-KIAC talk on 14th February 2025 (Friday).
  The details are as below:\n\n\n\nSpeaker: Thomas G Dietterich\, Universit
 y Distinguished Professor Emeritus\, School of Electrical Engineering and 
 Computer Science\, Oregon State University\nTitle: Can We Make Machine Lea
 rning Safe for Safety-Critical Systems?\nDate and Time: February 14\, 2025
 \; 04:00 to 05:00 PM (lecture) \; 05:15 to 05:45 PM (High tea and informal
  discussions with lecture attendees)\nVenue: #102\, CDS Seminar Hall.\n\n\
 n\nAbstract: The impressive new capabilities of systems created using dee
 p learning are encouraging engineers to apply these techniques in safety-c
 ritical applications such as medicine\, aeronautics\, and self-driving car
 s. This talk will discuss the ways that machine learning methodologies are
  changing to operate in safety-critical systems. These changes include (a)
  building high-fidelity simulators for the domain\, (b) adversarial collec
 tion of training data to ensure coverage of the so-called operational desi
 gn domain (ODD) and\, specifically\, the hazardous regions within the ODD\
 , (c) methods for verifying that the fitted models generalise well\, and (
 d) methods for estimating the probability of harms in normal operation. Th
 ere are many research challenges to achieving these.\n\nBut we must do mor
 e\, because traditional safety engineering only addresses the known hazard
 s. We must design our systems to detect novel hazards as well. We adopt Le
 veson’s view of safety as an ongoing hierarchical control problem in whi
 ch controls are put in place to stabilise the system against disturbances.
  Disturbances include novel hazards but also management changes such as bu
 dget cuts\, staff turnover\, novel regulations\, and so on. Traditionally\
 , it has been the human operators and managers who have provided these sta
 bilising controls. Are there ways that artificial intelligence (AI) method
 s\, such as novelty detection\, near-miss detection\, diagnosis and repair
 \, can be applied to help the human organisation manage these disturbances
  and maintain system safety?\n\nBio of Speaker: Thomas G Dietterich is Un
 iversity Distinguished Professor Emeritus in the School of Electrical Engi
 neering and Computer Science at Oregon State University. Dietterich is one
  of the pioneers of the field of machine learning and has authored more th
 an 220 refereed publications and two books. His current research topics in
 clude robust artificial intelligence\, robust human-AI systems\, and appli
 cations in sustainability.\n\nDietterich is the 2025 recipient of the Feig
 enbaum Prize for Applied AI and the 2024 recipient of the IJCAI Award for 
 Research Excellence. Dietterich is also the recipient of the 2022 AAAI Dis
 tinguished Service Award and the 2020 ACML Distinguished Contribution Awar
 d\, both recognising his many years of service to the research community. 
 He is a former President of the Association for the Advancement of Artific
 ial Intelligence and the Founding President of the International Machine L
 earning Society. Other major roles include Executive Editor of the journal
  Machine Learning\, Co-founder of the Journal for Machine Learning Researc
 h\, and Program Chair of AAAI 1990 and NIPS 2000. He currently chairs the 
 Computer Science Section of arXiv.org.\n\nHost Faculty: Prof. Jayant R Har
 itsa\n\n\n\nALL ARE WELCOME
CATEGORIES:Events,Talks
END:VEVENT
BEGIN:VTIMEZONE
TZID:Asia/Kolkata
X-LIC-LOCATION:Asia/Kolkata
BEGIN:STANDARD
DTSTART:20240215T160000
TZOFFSETFROM:+0530
TZOFFSETTO:+0530
TZNAME:IST
END:STANDARD
END:VTIMEZONE
END:VCALENDAR