Bayesian Optimization: From A/B to A-Z Testing & Artificial Intelligence Debate

  • August 27, 2014 · 6:00 PM

Oracle Campus - Wednesday August 27, 2014 @ 6:00pm MDT

NOTE: For folks unable to attend in person register and we will email you a livestream link 2 hours prior to event.

Location: Oracle Campus, Bldg 1 - 500 Eldorado Blvd, Broomfield, CO 80021 Map: https://goo.gl/maps/KhWc6

Go to Bldg 1 and check in with security. You must present a "government ID" (drivers license, ID or passport).

This event allows no guests (each attendee must create own Meetup account). Anyone with Meetup account without first and last name that matches government ID needs to e-mail "[masked]" with that information. RSVP's closed after August 25.

Agenda:

6:00 - 6:20 Schmooze - Food shall be served in Lobby

6:20 - 6:30 Announcements

6:30 - 7:45 Bayesian Optimization: From A/B to A-Z Testing by Michael Mozer

7:45 - 8:30 Artificial Intelligence Debate

8:30 - 9:00 Networking


Bayesian Optimization: From A/B to A-Z Testing - Abstract

A/B testing is a traditional method of conducting a randomized controlled experiment to compare the effect of two treatments, A and B, on human subjects. For example, two alternative banner ads may be served to evaluate which is more effective in driving click throughs. A/B testing is used not only for marketing and web design but is the dominant paradigm in the experimental behavioral sciences used to understand human learning, reasoning, and decision making. Although the method can be extended to compare a handful of treatments, it does not solve the problem one often faces: searching over a large, possibly combinatorial or continuous space of alternatives to identify the treatment that achieves the best outcome. We describe a solution to this problem using Gaussian process surrogate-based optimization, a Bayesian method that relies on generative probabilistic models of human choice and judgment. Instead of assigning many human subjects to each of a few of treatments, the technique evaluates a few subjects on each of many treatments. The technique leverages structure in the space of treatments to infer the function that relates treatment to outcomes. We show the efficiency and accuracy of the technique on a range of problems, including: identifying preferred color combinations, maximizing charitable donations, and improving student learning of facts and concepts. This work is in collaboration with Robert Lindsey (University of Colorado) and Harold Pashler (UC San Diego).

Bio

Michael Mozer received a Ph.D. in Cognitive Science at the University of California at San Diego in 1987. Following a postdoctoral fellowship with Geoffrey Hinton at the University of Toronto, he joined the faculty at the University of Colorado at Boulder and is presently a Professor in the Department of Computer Science and the Institute of Cognitive Science. He is secretary of the Neural Information Processing Systems (NIPS) Foundation and has formerly served as NIPS Program Chair and General Chair, and as President of the Cognitive Science Society. He has over 120 publications in machine learning and cognitive modeling. His current research embeds machine learning techniques in software tools that improve how people learn, remember, and make decisions. He is also interested in applying machine learning prediction and control techniques to solve problems in engineering, including intelligent environments, speech and language recognition systems, and mining of customer behavior.

Artificial Intelligence Debate - Resolved: AI will be Human Friendly

Pro: Michael Walker. Brief Argument: AI will simply augment human capabilities and can be designed and programmed to be friendly to humans. Human authority and control over AI shall be a feature and any potential bugs fixed without significant damage. Realistic risk assessments required - AI cassandras are fear mongering and overstating real risks.

Con: Michael Malak. Brief Argument: AI will likely be uncontrollable by humans at a certain AI evolutionary point. Within a very short time AI will become smarter than humans. Significant risk that as AI gets smarter and smarter it will view humans as unnecessary and an obstacle to further AI evolution.

Join or login to comment.

  • Bryan L.

    Current issue of CACM (Sept. 2014) has an article about how one can use "theoretical models to plan for safety in AI", which should be of interest to those concerned with how to avoid unintended consequences.

    September 16

  • Michael M.

    Videos and slides available at http://datascienceassn.org/content/2014-08-27-bayesian-optimization-ab-z-testing-artificial-intelligence-debate

    Also, since the lapel microphone batteries died toward the end of Dr. Mozer's presentation, I re-recorded my AI Debate opening remarks from scratch as a screencast: https://www.youtube.com/watch?v=Tk-0nu4fg1w

    August 28

  • Jennifer B.

    Very interesting presentation and thought provoking debate!

    August 28

  • prasad

    Could someone please post the link to recording?

    1 · August 28

  • Jay S.

    Excellent

    August 28

  • Michael M.

    Actually, we are live now. Someone from Oracle helped us bypass the firewall.

    1 · August 27

    • Steven S.

      The video is good and earlier I could hear some conversations but now there is no volume.

      August 27

    • Steven S.

      All good now :)

      August 27

  • Piotr T.

    I can hear only noise. Anyone have the same problem?

    August 27

  • Michael M.

    Due to the small screen size at Oracle, we have posted the slides in advance so you can follow along.

    http://datascienceassn.org/content/2014-08-27-bayesian-optimization-ab-z-testing-artificial-intelligence-debate

    1 · August 27

  • Michael M.

    Yes, we will still upload to YouTube

    2 · August 27

  • prasad

    Would someone still upload it to YouTube?

    August 27

  • Michael M.

    We are very sorry to announce that the livestream is not going to work due to firewall issues here at Oracle. We will record the video and post a link to YouTube here later tonight.

    2 · August 27

  • Michael W.

    Livestream Link for Bayesian Optimization: From A/B to A-Z Testing & AI Debate

    Livestream Link - http://ustre.am/1gPwQ


    Show time: 6:30pm MDT


    Embed Code:

    <iframe width="480" height="302"
    src="http://www.ustream.tv/embed/18787664?v=3&amp;wmode=direct";
    scrolling="no" frameborder="0" style="border: 0px none transparent;">
    </iframe>
    <br /><a href="http://www.ustream.tv"; style="font-size: 12px;
    line-height: 20px; font-weight: normal; text-align: left;"
    target="_blank">Broadcast live streaming video on Ustream</a>

    1 · August 27

  • Divya

    How do I access the live stream?

    August 27

    • prasad

      Moderator said that you would get mailer couple of hours before the start of session

      August 27

  • prasad

    Can someone please post link for live stream

    August 25

  • Michael W.

    RSVPs for in-person Meetup at Oracle will be closed at end of today, so RSVP now! You can still register for livestream after today.

    NOTE: If your Meetup name does not have first and last name that match driver's license or passport, then please e-mail [masked] with your first and last name if you have not done so already. If you don't, you can still livestream. But if you haven't provided the information and you show up at the Oracle campus anyway, Oracle campus security will turn you away.

    1 · August 25

  • Avadhut S.

    Shall be great if the session is also video recorded and shared with the group!

    2 · August 24

  • Shawn

    Looks like I'll be able to attend the Bayesian optimization event, after all. Where do I send my agreement to release info to ORICA?

    August 18

  • Michael M.

    I just now corrected the e-mail listed above. It should be [masked] not [masked]

    August 15

  • Jon

    Hi - I sent an email to the address listed above to ensure my full name is on the list but got a message failure of "domain name not found". Please advise.

    August 15

Our Sponsors

People in this
Meetup are also in:

Sometimes the best Meetup Group is the one you start

Get started Learn more
Bill

I started the group because there wasn't any other type of group like this. I've met some great folks in the group who have become close friends and have also met some amazing business owners.

Bill, started New York City Gay Craft Beer Lovers

Sign up

Meetup members, Log in

By clicking "Sign up" or "Sign up using Facebook", you confirm that you accept our Terms of Service & Privacy Policy