Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[sea-list] AI Safety at OpenAI (Today) at 5:30pm 4/24 @ Varian Physics 355

Jake McKinnon jakemck at stanford.edu
Wed Apr 24 08:46:03 PDT 2019


You are invited to the weekly transformative* AI safety working group
<https://nthomas103.github.io/stais/>.

This week, OpenAI <https://openai.com/> researcher Geoffrey Irving
<https://naml.us/>will lead a discussion about safety work at OpenAI,
emphasizing human reward learning <https://arxiv.org/abs/1706.03741>, safety
via debate <https://arxiv.org/abs/1805.00899>, and why AI safety needs
social scientists <https://distill.pub/2019/safety-needs-social-scientists>.

This event is targeted at graduate students, but highly interested
undergraduates (especially those involved in Stanford EA, since we're
helping run this) are welcome as well!

Details:  5:30pm Wednesday, April 24th in Varian physics 355 (3rd floor).
Email Nate Thomas (ncthomas at stanford.edu) with your questions.

Hope to see you there!

*By "transformative AI", we mean that we'll mostly be focusing on long-term
challenges related to "AI that precipitates a transition comparable to (or
more significant than) the agricultural or industrial revolution" (from the
Open Philanthropy Project blog
<https://www.openphilanthropy.org/blog/some-background-our-views-regarding-advanced-artificial-intelligence#Sec1>
)" in this working group. (Though we're really excited to see expanding
efforts in related areas too!)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/think-list/attachments/20190424/6d4d1d2f/attachment.html>


More information about the think-list mailing list