5 replies [Last post]
beta_helix's picture
User offline. Last seen 1 day 7 hours ago. Offline
Joined: 05/09/2008
Groups: None

I wanted to share with all of you an hour-long public lecture that I gave at the Institute for Computational and Experimental Research in Mathematics (ICERM) at Brown University in April 2018:

Joined: 09/24/2012
Groups: Go Science

Thanks for the share and the respect you show for players.

What is interesting is that we have the impression, listening, that we share this story with you ;-)

Joined: 05/19/2017
Groups: None
Thank you for sharing!

Seeing the evolution of Foldit's design (and why some tutorials are in the game in the first place despite being out of practice with regards to the weekly Science Puzzle rotations) was really cool and I'm glad you had the opportunity to share it like this!

And hey, if you wanna learn electron density puzzles, I'm sure we'd have your back!

Susume's picture
User offline. Last seen 1 hour 18 min ago. Offline
Joined: 10/02/2011
Thanks for the link!

Very enjoyable lecture!

Were the foldit results from the ED puzzles you talk about (741b, 746, 750) ever published anywhere? Players beating automated systems on blind problems seems like just what foldit was designed for, but I don't remember hearing about these results before.

beta_helix's picture
User offline. Last seen 1 day 7 hours ago. Offline
Joined: 05/09/2008
Groups: None
Great question!

We had started writing the manuscript for that paper, as they were indeed incredibly exciting results: https://fold.it/portal/node/995693

But we felt like it still needed a final case: either something unsolved (like the MPMV puzzle) or comparing against human methods (like a CASP experiment for Electron Density).

Then we were lucky enough to have our colleagues at the University of Michigan create exactly such a competition (excuse me, I meant experiment ;-) not only with automated methods and crystallographers, but also with Undergraduate Students participating: https://fold.it/portal/node/2002791
We did want to include those previous results in this paper, but we ran out of room. Even the Supplemental Material section became long enough already!

We actually ran into a similar issue with the first Foldit Nature Paper, where we had some exciting initial results (Figures S8-S10 on pages 16-18) but because they weren't blind, we couldn't include them in the main paper (and had to move them to the Supplemental Material).

These papers need to be so concise, that once you have an amazing result (such as the Foldit case in the Nature Communications paper) it becomes difficult to fit other exciting results that happened beforehand, even though they led up to the latest result. (That's why I like presenting lectures, where you can slowly build up the exciting results and end with a bang! :-)

Joined: 09/24/2012
Groups: Go Science
BTW, the evolution of player's strategies

In pages 28-45 of supplementary material of the 2010 Foldit Nature Paper, players explain their (mainly hand fold) strategies. These strategies are now partly embedded in recipes.
It's amazing how players's strategies changed over time, now that we have plenty of recipes to use with some focused hand fold.

These "old" stories might be compared to recent stories from page 15 of the supplemental material of the 2018
Human Computation paper.

Yet, supplementary materials are still shortcuts of the basic information. For the late paper, I recorded a 16 minutes annotated video only for 1.5 pages of supplementary material, and much more during one week for another part of the paper.

The fact is that there is no other way to develop a "professional vision" of the game than to play and play again (the learning curve is 2 years). And to take time to watch the Black Belt video series.


Developed by: UW Center for Game Science, UW Institute for Protein Design, Northeastern University, Vanderbilt University Meiler Lab, UC Davis
Supported by: DARPA, NSF, NIH, HHMI, Amazon, Microsoft, Adobe, Boehringer Ingelheim, RosettaCommons