Discover and read the best of Twitter Threads about #AlphaStar
Most recents (7)
Looking for a few more favourite resources from the team? Today’s #AtHomeWithAI picks are from software engineer Julian Schrittwieser (@Mononofu), one of the team behind #AlphaZero!(1/6) 
Many people are now moving code into Jax. To get up to speed, Julian suggests exploring the DeepMind Haiku library, designed to help you implement deep reinforcement learning algorithms. Explore it here: bit.ly/2zBIkpR #AtHomeWithAI
Fancy a long read? Though the field of DL moves very quickly, @Mononofu considers both good foundational resources:
Neural Networks and DL from @michael_nielsen bit.ly/3ePYa06
DL by @goodfellow_ian, Yoshua Bengio & @AaronCourville bit.ly/351qMzb #AtHomeWithAI
Neural Networks and DL from @michael_nielsen bit.ly/3ePYa06
DL by @goodfellow_ian, Yoshua Bengio & @AaronCourville bit.ly/351qMzb #AtHomeWithAI
After high school, I ended up working odd jobs, e.g., as a night shift cleaner at a fast-food restaurant. One especially nasty night, I decided that I had to at least try to get more out of my life and began self-studying coding every free minute I had, up to 20 hrs/day.
1/5
1/5
In March 2010, I landed my first programming job. Paid peanuts and involved legacy PHP4 codebase but meant the world that somebody took a chance on me. Five years later, I was a senior web dev at @KNPLabs with great bosses, mentors, and friends @weaverryan and @leannapelham.
2/5
2/5
@KNPLabs @weaverryan @leannapelham At some point, I decided to try and somehow contribute to something I was interested in since childhood -- AI, which in practice meant rebuilding my career almost from the ground up. Having virtually no maths background, I began with an undergrad in statistics at @unitartu.
3/5
3/5
Fabio Ciucci on LinkedIn [1/8]: #Putin said "the nation that leads in #AI will rule the world," not that "a rebel AI will rule the world starting with burning the nation that built such AI." #ArtificialIntelligence
Fabio Ciucci on LinkedIn [2/8]: If #Russia and #China lead on weaponized AI (and they could), other nations including the USA (whatever they deny) do it too, like with atomic bombs.
Fabio Ciucci on LinkedIn [3/8]: News: Google AI (#DeepMind #AlphaStar) finally became able to kill every pro-human at #starcraft2 war game too.
Today at 3am Tokyo time, @DeepMindAI presented one new member of the Alpha family: #AlphaStar, playing StarCraft 2.
#AlphaStar beat two SC2 pro gamers, @LiquidTLO and @Liquid_MaNa, and this, Ladies and Gentlemen, is a real achievement. So can we say that SC2 is beaten, like Go of Chess? Definitely no.
I actually watched the whole #AlphaStar demo this morning with my girlfriend, who unlike me actually plays StarCraft. Read the article (vox.com/future-perfect…), but thoughts:
The systems we watched were trained for seven and fourteen days real-time, which is 200-400 years of gameplay time accumulated. In a way, "number of days" is misleading as a stat, it's pretty much just a function of how much compute you bought.
Nonetheless I think it merits a mention, because ....DeepMind decided in November or December to focus here. They then got top-pro level play in the space of about a month real time. Yes, this is because they can do a lot in parallel. But... they can do a lot in parallel.
What happened with #AlphaStar? Checkout out @DeepMindAI’s official blog here: deepmind.com/blog/alphastar… . I’ll be unpacking some stuff based on the information from the live stream and now blog post.
Context: After Deep Blue beat a chess grandmaster, Go - with its considerably larger action space and lack of heuristics - seemed to be the next big challenge. Like AlphaStar, the original AlphaGo used a combination of imitation learning and DRL to learn to play.
AlphaZero replaced imitation learning with self-play, but like its predecessor used tree search for training by rolling out a Go simulator. SC II is far too expensive to use in this fashion, so search during play or even expert iteration in training wouldn't work.
1/3 Really proud to share #AlphaStar with the world - a milestone for game-playing AI, and the first system to beat a top pro at #StarCraft2. Massive thanks to @Blizzard_Ent, @Liquid_MaNa & @LiquidTLO for being part of this great journey!
Blog + Replays: deepmind.com/blog/alphastar…
Blog + Replays: deepmind.com/blog/alphastar…
2/3 As well as showing how far we’ve come, it was exciting to test out our new prototype system that uses the in-game camera interface to play. Massive respect to @Liquid_MaNa for outplaying it! We’ll keep working on it and play more games soon :)
3/3 While StarCraft is ‘just’ a (very complex!) game, I’m excited that the techniques behind #AlphaStar could be useful in other problems such as weather prediction & climate modeling, which also involve predictions over very long sequences. Peer-reviewed paper is underway.
