4 ms·

I'm also excited to see the future of this bot when they demonstrate a terran AI with near-perfect marine/stim/medivac micro.

kzrdude 8y ago
Perfect micro bots don't excite me much, because they've existed all along, and it's not an AI task.
- WJW 8y ago
  There was a "perfect zergling micro vs siege tanks" bot some time ago that would micro lings away from the one that was being fired at by the tanks, thereby negating all the splash damage. The effect was insanely powerful. But as you say, showing that a bot can have perfect micro is not very interesting. Of course a computer can have better control of well defined tasks like moving a unit away just before it dies, especially doing so for many different units concurrently. What is interesting is the wider strategy and how the computer deals with imperfect information.
- sjmulder 8y ago
  Here’s that perfect zergling video: https://youtu.be/IKVFZ28ybQs https://youtu.be/IKVFZ28ybQs
- mithr 8y ago
  The interesting part to me is that, as far as I understand, the AI figured out this strategy by itself, basically deciding that it would be a good way for it to win games, rather than being specifically programmed to do it. That's actually pretty cool! Other than that, I agree, and am also much more interested in what happens when you have a more level playing field (using camera movement rather than API, limiting reaction times and CPM, etc). I look forward to future matches where this happens.
  - omgtehlion 8y ago
    It was not programmed to do the thing, but all these tactics were in seed replays, from which the agent started its learning. So, it actually not figured the move _by itself_, only found it useful.
  - vladimirralev 8y ago
    I think there is some debate about what the neural net did and what was hardcoded. So far all starcraft AIs consist of hardcoded intelligent micro ruled by a neural net that picks one out of less than 100 possible hardcoded choices. And things like "expand", "scout", "group units", "micro" are hardcoded outside of the neural net, part of the API in fact. When the researches said they only used 15 TPUs for 14 days on LSTM, this makes me think they really narrowed down the search space of the neural net and hardcoded a lot of the micro or at least trained separate micro nets.
  - richardknop 8y ago
    Not really. The version which learned from scratch was scrapped as it didn’t work at all. This version learned by observing pros. So it didn’t learn by itself, it imitated and perfected pro players.
hughzhang 8y ago
I'm scared. Medivacs only healing the front line and perfect stimming only the backline will be SOOO broken.

kzrdude 8y ago

Perfect micro bots don't excite me much, because they've existed all along, and it's not an AI task.

WJW 8y ago
There was a "perfect zergling micro vs siege tanks" bot some time ago that would micro lings away from the one that was being fired at by the tanks, thereby negating all the splash damage. The effect was insanely powerful. But as you say, showing that a bot can have perfect micro is not very interesting. Of course a computer can have better control of well defined tasks like moving a unit away just before it dies, especially doing so for many different units concurrently. What is interesting is the wider strategy and how the computer deals with imperfect information.
sjmulder 8y ago
Here’s that perfect zergling video: https://youtu.be/IKVFZ28ybQs https://youtu.be/IKVFZ28ybQs
mithr 8y ago
The interesting part to me is that, as far as I understand, the AI figured out this strategy by itself, basically deciding that it would be a good way for it to win games, rather than being specifically programmed to do it. That's actually pretty cool! Other than that, I agree, and am also much more interested in what happens when you have a more level playing field (using camera movement rather than API, limiting reaction times and CPM, etc). I look forward to future matches where this happens.
- omgtehlion 8y ago
  It was not programmed to do the thing, but all these tactics were in seed replays, from which the agent started its learning. So, it actually not figured the move _by itself_, only found it useful.
- vladimirralev 8y ago
  I think there is some debate about what the neural net did and what was hardcoded. So far all starcraft AIs consist of hardcoded intelligent micro ruled by a neural net that picks one out of less than 100 possible hardcoded choices. And things like "expand", "scout", "group units", "micro" are hardcoded outside of the neural net, part of the API in fact. When the researches said they only used 15 TPUs for 14 days on LSTM, this makes me think they really narrowed down the search space of the neural net and hardcoded a lot of the micro or at least trained separate micro nets.
- richardknop 8y ago
  Not really. The version which learned from scratch was scrapped as it didn’t work at all. This version learned by observing pros. So it didn’t learn by itself, it imitated and perfected pro players.

hughzhang 8y ago

I'm scared. Medivacs only healing the front line and perfect stimming only the backline will be SOOO broken.