BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Motivation for this group\, Goodhart's Law - James Bell\, Universi
 ty of Cambridge
DTSTART:20181017T160000Z
DTEND:20181017T173000Z
UID:TALK113134@talks.cam.ac.uk
CONTACT:Adrià Garriga Alonso
DESCRIPTION:How can we design AI systems that reliably act according to th
 e true intent of their users\, even as the capability of the systems incre
 ases?\n\nCome to this reading group with *free pizza*! This week we will g
 et started by motivating why we are doing this. In part\, this is \nGoodha
 rt's Law [1] and its implications for evaluating AI systems\, and designin
 g their objectives.\n\nThe session will go as follows. At 17:00\, we will 
 start reading the material (see bottom)\, mostly individually. At 17:30\, 
 the discussion leader will start going through the paper\, making sure eve
 ryone understands\, and encouraging discussion about its contents and impl
 ications.\n\nA basic understanding of machine learning is helpful\, but de
 tailed knowledge of the latest techniques is not required. Each session wi
 ll have a brief recap of immediate necessary knowledge. The goal of this s
 eries is to get people to know more about the existing work in AI research
 \, and eventually contribute to the field.\n\nJoin the mailing list (https
 ://lists.cam.ac.uk/mailman/listinfo/eng-safe-ai)\, the Facebook group (htt
 ps://www.facebook.com/groups/1070763633063871) or the talks.cam page (http
 s://talks.cam.ac.uk/show/index/80932). Announcements about the week's topi
 c and other events will be sent there. Consider also inviting your friends
 !\n\nREADING MATERIAL:\n\n"Building safe artificial intelligence: specific
 ation\, robustness\, and assurance" (2018)\, by Pedro A. Ortega\, Vishal M
 aini\, and the DeepMind safety team https://medium.com/@deepmindsafetyrese
 arch/building-safe-artificial-intelligence-52f5f75058f1\n\n"On the folly o
 f rewarding A\, while hoping for B" (1975)\, by Steven Kerr http://web.mit
 .edu/curhan/www/docs/Articles/15341_Readings/Motivation/Kerr_Folly_of_rewa
 rding_A_while_hoping_for_B.pdf\n\n"Categorizing Variants of Goodhart's Law
 " (2018)\, by David Manheim and Scott Garrabrant (arXiv https://arxiv.org/
 abs/1803.04585)\n\nIf you have already read the material in your own time\
 , feel free to come by at 17:30.\n\n[1] https://en.wikipedia.org/wiki/Good
 hart%2527s_law\n\n
LOCATION: Cambridge University Engineering Department\, CBL Seminar room B
 E4-38.  See https://www.openstreetmap.org/#map=18/52.19804/0.11969
END:VEVENT
END:VCALENDAR