Adam Stelmaszczyk

[Adam Stelmaszczyk @astelma1]

Started a blog on https://t.co/972rmJ8WX8. Imported 3 posts from Medium + just finished 1 consolidating @FakePsyho's tips from https://t.co/jTO0Z2ngqj. Enjoy!

[Psyho @FakePsyho]

Let's make a silly experiment.
For every like this post receives in the next 24 hours, I will post one tip for heuristic/bot programming contests.

I will start posting these as a thread in ~1h from now.

I could talk about those things for days, so bring it on!

[Adam Stelmaszczyk @astelma1]

@FakePsyho Woah, thanks a lot for all the sharings, great stuff. Did I get the idea of Fast-Clearing Array right? https://t.co/BR3KwfvbQf

[Adam Stelmaszczyk @astelma1]

@unixpickle If you'd like to try more cards, you could probably exclude those expensive black cards.

Another thought: pro agent (when being a judge) would try to avoid awarding a point to his biggest rival in terms of points (faking that somebody else was the funniest).

[Adam Stelmaszczyk @astelma1]

@unixpickle Interesting. Perhaps varying the number of white cards could also shed more light; more cards, more expressive (...to some saturation point?). Wiki mentions 10, the code uses 7.

[Adam Stelmaszczyk @astelma1]

Just shipped "4th place solution to Python Speed Coding Challenge" https://t.co/K8p0h3Dv56

[Adam Stelmaszczyk @astelma1]

@DaveChurchill It can, it's easy to do some compiler optimization slightly differently, impacting the time when testing with one particular code. Another example: https://t.co/ug9euaxbnP

[Adam Stelmaszczyk @astelma1]

@DaveChurchill Interesting, this made me consider to try it for https://t.co/K6pIWbVhaA. What speedup a custom stack gave in your case?

[Adam Stelmaszczyk @astelma1]

@Inoryy Yeah, also producing a bit more workers (could die from harassment, could move them to expansion) https://t.co/n7vvqXuoM6

[Adam Stelmaszczyk @astelma1]

@DaveChurchill https://t.co/7HxVjnnRrx

[Adam Stelmaszczyk @astelma1]

@json_dirs Maybe @lukaszkaiser knows.

[Adam Stelmaszczyk @astelma1]

@unixpickle Great, congrats! What's happening on the floor 20, some new obstacle? Do you know why your approach can't handle that? Do you have some ideas how to improve it?

[Adam Stelmaszczyk @astelma1]

@Inoryy @DeepMindAI Awesome, congrats! 🎉

[Adam Stelmaszczyk @astelma1]

@pretendsmarts @aicrowdHQ @unity3d I think keys unlocking doors are introduced.

[Adam Stelmaszczyk @astelma1]

@jonasschneider What behaviors are scripted?
What would you do differently if you could redo Five?
What were the most surprising things during its development?
What are the plans for the future?

Exciting work, well done :)

[Adam Stelmaszczyk @astelma1]

@awjuliani @pyoudeyer Maybe: wavy water, grass, leaves, trees, animals, rain, drizzle, fog, changes in lighting etc. :)

[Adam Stelmaszczyk @astelma1]

I highly recommend this Machine Learning School!

[EEML @EEMLcommunity]

We are proud to announce the 2019 edition of EEML summer school, 1-6 July, Bucharest, Romania. Topics covered: DL, RL, computer vision, bayesian learning, medical imaging, and NLP. An amazing set of speakers confirmed so far! More info coming soon! Check https://t.co/PM3Tn3ox4Y!

[Adam Stelmaszczyk @astelma1]

@woj_zaremba I'd say yes, if backprop would have the same exploration as evolutionary algorithms. They are very similar to backprop / gradient methods, population follows the gradient without explicitly calculating it.

[Adam Stelmaszczyk @astelma1]

@FakePsyho What did you need the transcript for? Other than a bit of twitter fun ;)

[Adam Stelmaszczyk @astelma1]

@tambetm I've read the article, found it very interesting and then realized you are the author, great piece :) Also thanks for the reference.

[Adam Stelmaszczyk @astelma1]

@FakePsyho @isenbaev @jpsbur @HaliteAI @CodinGame You can also check out @crowd_ai, they usually host several DS/ML/RL competitions