StarCraft II: A New Challenge for Reinforcement Learning · HackerPapers discussions

StarCraft II: A New Challenge for Reinforcement Learning — mentioned by runnerup

Wed, 19 Jan 2022 14:48:12 GMT

Edit 2: Reading through the "supplementary data" of the 2019 paper, it definitely appears that the AlphaStar which reached grandmaster was not limited in the same ways as the 2017 paper would suggest. x/y positions of units are not determined visually, but fed directly from the API. So AlphaStar absolutely can just run Attack(Position: carrier_of_interest->pos.x) and not mis-click. It's "map" / "vision" is really just a bounding box of an array of every entity/unit on the map and all the things that a human would have to spend APM to manually check (precise energy level, precise hit points…

Paper and discussion history. A mention is not an endorsement.

StarCraft II: A New Challenge for Reinforcement Learning — mentioned by gwern

Tue, 22 Jan 2019 23:50:18 GMT

180 APM, specifically: pg6 https://arxiv.org/abs/1708.04782

Paper and discussion history. A mention is not an endorsement.

StarCraft II: A New Challenge for Reinforcement Learning — HN submission

Thu, 17 Aug 2017 08:08:16 GMT

StarCraft II: A New Challenge for Reinforcement Learning

3 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.