Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Comment author: waveman 11 July 2017 01:35:59AM 1 point [-]

Can I suggest you put the name of the city at the start of the title, so that everyone doesn't need to go into the posting to check is it their city?

Others do this. Thanks,

Comment author: MathieuRoy 27 July 2017 07:53:21AM 0 points [-]

Just saw your comment. Thanks for letting me know.

Meetup : Montreal - AI Safety: Orthogonality & Convergent instrumental goals (MIRI's theses part 1)

0 MathieuRoy 24 July 2017 06:26AM

Discussion article for the meetup : Montreal - AI Safety: Orthogonality & Convergent instrumental goals (MIRI's theses part 1)

WHEN: 27 July 2017 07:00:00PM (-0400)

WHERE: room 3195, Pavillon André-Aisenstadt, Université de Montréal

Discussion article for the meetup : Montreal - AI Safety: Orthogonality & Convergent instrumental goals (MIRI's theses part 1)

Meetup : AI Ethics: Deep dive on how to embed values into machines

0 MathieuRoy 11 July 2017 06:00PM

Discussion article for the meetup : AI Ethics: Deep dive on how to embed values into machines

WHEN: 19 July 2017 07:00:00AM (-0400)

WHERE: 2920 Ch de la Tour, Montréal, QC

Meetup event for all information: https://goo.gl/1pzC2z Exact location: room 3195, Pavillon André-Aisenstadt, Université de Montréal

Summary: This is the second session in our bi-weekly AI Ethics group where we will be making a deep dive into how best to embed values into machines, what the implications of that are and concrete proposals that we can have around it.

Discussion article for the meetup : AI Ethics: Deep dive on how to embed values into machines

Meetup : AI Safety: Reinforcement Learning with a Corrupted Reward Channel

0 MathieuRoy 10 July 2017 06:56PM

Discussion article for the meetup : AI Safety: Reinforcement Learning with a Corrupted Reward Channel

WHEN: 10 July 2017 07:00:00PM (-0400)

WHERE: 2920 Ch de la Tour, Montréal, QC

Location: room 3195, Pavillon André-Aisenstadt, Université de Montréal

Contact me, Mati, at 438-886-0994 if you can't find us.

The paper is available here: https://arxiv.org/abs/1705.08417

This is a recent collaboration between DeepMind and Australian National University.

Abstract: << No real-world reward function is perfect. Sensory errors and software bugs may result in RL agents observing higher (or lower) rewards than they should. For example, a reinforcement learning agent may prefer states where a sensory error gives it the maximum reward, but where the true reward is actually small. We formalise this problem as a generalised Markov Decision Problem called Corrupt Reward MDP. Traditional RL methods fare poorly in CRMDPs, even under strong simplifying assumptions and when trying to compensate for the possibly corrupt rewards. Two ways around the problem are investigated. First, by giving the agent richer data, such as in inverse reinforcement learning and semi-supervised reinforcement learning, reward corruption stemming from systematic sensory errors may sometimes be completely managed. Second, by using randomisation to blunt the agent's optimisation, reward corruption can be partially managed under some assumptions. >>

Also see Tom's blog post about this paper here: http://www.tomeveritt.se/paper/2017/05/29/reinforcement-learning-with-corrupted-reward-channel.html

Please do your best to read the paper in advance, so we can have a good discussion about the content.

If necessary, David Krueger will give a run-down of the paper.

Discussion article for the meetup : AI Safety: Reinforcement Learning with a Corrupted Reward Channel

Meetup : AI Ethics: Planning and Introductions

0 MathieuRoy 01 July 2017 06:28AM

Discussion article for the meetup : AI Ethics: Planning and Introductions

WHEN: 03 July 2017 01:00:00PM (-0400)

WHERE: 2900 Edouard Montpetit Blvd, Montréal, QC

If you're in Montreal and are interested in AI Ethics and / or AI Safety, I encourage you join the Montreal AI Ethics' Meetup group: https://goo.gl/Z8jZPm Our next event will be on AI Ethics, but there will be events on AI Safety as well. The official description is below.


The first AI Ethics reading group meeting will be a discussion group at lunch time on Monday!

Where: P-310, Pavillion Roger Gaudry on UdeM campus (main room of the Deep Learning Summer School)

When: Monday July 3, 13:00-14:30 (lunch time of the DLSS),

Topics up for discussion under 'AI Ethics' include (but are not limited to!):

• ensuring the benefits of AI are democratically available

• identifying and working on problems of social importance

• preventing weaponization and/or malicious use of AI

• quantification and correction of bias in datasets

• data ethics and privacy

• identifying the roles of researchers, policy makers, end-users, and other groups; what concrete actions each should be taking

The schedule for the discussion will be as follows:

13:00 - 13:10 (5-10 mins): Brief introduction about the reading group 13:10 - 13:30 (15-20 mins): Initial list of topics proposed by organizers, refined by audience 13:30 - 14:00 (30 mins): Split into smaller discussion groups to develop concrete, feasible proposals 14:00 - 14:20 (20 mins): Brief presentation of each group's thoughts and proposals by a spokesperson

This is not a 'deep-learning exclusive' group; members of the community and anyone interested in these issues are warmly invited to participate.

If you have any questions or difficulty finding the place, you may contact Mati at 438-886-0994 or montreal@eahub.org.

Discussion article for the meetup : AI Ethics: Planning and Introductions

Meetup : Sunrise on top of Mont Royal: Argument (5/12)

0 MathieuRoy 09 June 2017 03:15AM

Discussion article for the meetup : Sunrise on top of Mont Royal: Argument (5/12)

WHEN: 12 June 2017 05:00:00AM (-0400)

WHERE: Chalet Mont Royal, Montreal

Reference: https://www.facebook.com/events/303698656754295/

We're leaving the Macroscope (1191 Avenue Hope #1) at 4:30 in order to be on top of Mont Royal for the sunrise which is at 5:00. You can either join us at the Macroscope or at the top.

A short text on Argument, one of the 12 virtues of rationality, will be read once at the top of Mont Royal.

This can be a good opportunity to put your coming week into perspective.

For any questions, you may reach Mati at mathieu.roy.37@gmail.com.

Source of the texts: "Twelve Virtues of Rationality" by Eliezer Yudkowsky http://yudkowsky.net/rational/virtues/

Discussion article for the meetup : Sunrise on top of Mont Royal: Argument (5/12)

Meetup : Rationality Potluck

0 MathieuRoy 07 June 2017 03:53AM

Discussion article for the meetup : Rationality Potluck

WHEN: 21 June 2017 07:00:00PM (-0400)

WHERE: 2-1191 Avenue Hope, Montreal, Qc, H3H2A3

Facebook Event Reference: https://goo.gl/1YB7ZK

You're welcome to come share food / beverage with other rationality enthusiasts, and celebrate the Summer Solstice: the longest day of the year!

At some point during the evening, Marko Thiel, a researcher in pure mathematics, and an organizer of LessWrong and MIRIx events in Zurich, Switzerland, will make a brief presentation about epistemic independence versus herding (with election polls as the example).

Discussion article for the meetup : Rationality Potluck

Meetup : Rationality Potluck

0 MathieuRoy 25 May 2017 06:28PM

Discussion article for the meetup : Rationality Potluck

WHEN: 25 May 2017 06:30:00PM (-0400)

WHERE: 1191 Avenue Hope, Montreal

Eric Chisholm from the Vancouver Rationalist Community is staying at the Macroscope this week. You're invited to come say hi, and talk with other rationality enthusiasts! Feel free to invite friends.

Bring food and/or beverage if possible. Vegan food will be available.

Eric Chisholm, an alumnus from the Center for Applied Rationality, will present the Double Crux technique: a technique for resolving disagreement.

Facebook Event : https://goo.gl/f8Uwfg

Discussion article for the meetup : Rationality Potluck

[Link] Wikipedia book based on betterhumans' article on cognitive biases

1 MathieuRoy 14 October 2016 01:03AM
Comment author: MathieuRoy 17 January 2015 05:33:04AM *  0 points [-]
  • jumpsuits/onepieces (I find them really comfortable)
  • if you don't have a lot of dishes (ex.: live alone), something like this to avoid putting your hands in hot water, and with soap in the handle to be more efficient
  • a second pillow to put between or below your legs when you sleep

View more: Next