Sunday Biweekly meetup: Occam's Razor, Repetition and Time's Up

[John Salvatier]

Hi All!

The biweekly meetup is happening at the regular place and time. The plan is to:

• Read and then discuss EY's post Occam's Razor
• Play a variant of Repetition (two people conversing, one making observations)
• Play the good version of Time's Up

What:   regular LW meetup 
Date:   Sunday the 18th
Time:   4pm - 6:30pm 
Where: 414 50th St NE, Seattle, WA

We will be ordering take out again, so please bring some cash if you expect to be hungry.

The official end of the meetup will be 6:30pm, but people are welcome to stay and hang out after that.

The week after this, Guy will be doing a tutorial on Bayes theorem which should be fun.

Looking forward to seeing you guys!

[John Salvatier]

Hi All, 

I think the Occam's razor discussion worked well, though we determined that Solomonoff induction could definitely use good layman's introduction.   The repetition variant didn't work very well and several people said they didn't really get much out of playing repetition. Time's up was a fun game. 

Here are some things I promised to link to:

Someone mentioned they hadn't read That Alien Message, which I think is a really good EY short story. 

During dinner custom fitted shirts were mentioned such as blanklabel. It's also the case that many of these services have introductory deals ($20 shirt).

Brendan mentioned he was looking to learn statistics for which I recommend part 1 of Data Analysis - A Bayesian Tutorial (download) which is about 100 pages.

See you guys next week for Guy's tutorial on bayes' theorem!

[Brandon Reinhart]

For statistics, I recommend An Introduction to Bayesian Statistics by William Bolstad. This is superior to the "Data Analysis" book if you're learning stats from scratch. Both "Data Analysis" and "Bayesian Data Analysis" assume a certain base level of familiarity with the material. The Bolstad book will bootstrap you from almost no familiarity with stats through fairly clear explanations and good supporting exercises.

Nonetheless, it's something you should do with other people. You may not notice what you aren't completely comprehending otherwise. Do the exercises!

=Brandon

--
Brandon Reinhart

Valve Software

[John Salvatier]

Ok, good to know that! I actually started with Bolstad's book but later thought the Data Analysis book was better afterwards, so maybe that's just me  underestimating inferential distance. The exercises are key.

I thought there was a fair amount of not so useful things in Bolstad's book comparing the Bayesian and frequentist approaches. 

There's a fair number of people planning on learning a math topic in the near future. I wonder if we should do a math study group. 

[Brandon Reinhart]

Yeah, Bolstad's book contains some deprogramming for people who had some exposure to frequentism. I don't mind this, but in my group are two psychologists who are trained in frequentism, so the comparisons are useful.

Getting both books doesn't hurt.

My group is doing Calculus on Khan Academy as well, since that will be needed at some point. (And most of us are really rusty...)

=Brandon

[John Salvatier]

So you have a regular math study group?

[Brandon Reinhart]

Yeah only four people including myself. We all work at the same place, so we meet Wednesdays at 11 for an hour or two, do exercises, talk about concepts, work the whiteboard, stuff like that. We're less interested in proofs and more identifying things we can use at work.

=Brandon

[John Salvatier]

Very cool.

[Morgan Catha]

I agree that is cool.

[John Salvatier]

I found an electronic copy of Bolstad's book (dropbox).