![]() ![]() On top of these challenges real-time strategy games due their continuous time and space add another layer of complexity (e.g., optimal formation of units, evade balistic projectiles or missiles). Hence, "naive" greedy approaches to improving a single score (e.g., Flipper, Super Mario, etc.) won't cut it. The longer an effect takes to materialize, mapping the actions/decisions to the very effect gets more complicated. needs to be more dynamic and will face more sparce option spacesĬ) More long-term consequences: In many cases, effect of actions and decisions can only be seen after several rounds, minutes of play. The key challenges when switching from board games to strategy games is:Ī) Knowledge of state of board is incomplete - therefore information has a cost: Due to fog of war (and unclear worklists, resources) position/status of the enemy and overall map is not fully known - hence risky investments (explorer teams) are required to gain informationī) Number of actors/actions is unlimited: Buildings, Cities, Units can be produced infinitely - hence the data model and learning method for neural networks etc. ![]() OpenAI and Deepmind already start tackling more complex strategy games like Dota 2 and Starcraft 2. Recent breakthroughs by AlphaZero, AlphaGo show capabilities of current systems to master "simple" board games via reinforcement learning and Montecarlo tree search (MCTS). ![]() Looking for contributors, AI and Freeciv enthusiasts. The main goal is to foster AI research using a round-based strategy game like freeciv. they play more like board games requiring considerably less micro management).Freeciv-bot allows to steer freeciv via Python and build automated bots that can ultimately beat human players. Right now, AlphaStar is able to beat virtually every human SCII player (even Grandmasters) and I bet that world conquest in EUV would not be casual gaming anymore if AlphaStar were on the map.įinally, I would be interested in how DeepMind's reinforcement learning would work for Paradox games in general since they are, compared to SCII, more like Go (i.e. If Paradox would offer a DLC with AlphaStar agents - even though experimentally - I would definitely give it a try since it would make the game considerably more challenging in single player mode. Thus I am suggesting to limit such an AI to very specific countries similar to the initial training of AlphaStar (only one map, and only one race). However, the AlphaStar AI might not work for every single country tag in the game since it will require more processing power than usual AI (not to mention the capacities you need to train the agents, but that would be outsourced anyways). Indeed, that would be an entirely different discussion (and I am not insisting on keeping this concept). In case you are interested in AlphaStar, there is an article published in Nature Magazine and two articles by DeepMind.Īnd of course, something like AlphaStar would address those problems in current games which are linked to AI behavior (which is understandable since this is probably one of the most difficult aspects of development). ![]() However, I believe that DeepMind demonstrated that learning machines might be the future of AI in video games and AlphaStar is also highly adaptable (in fact, as I understood, it recently played the ladder starting with Bronze League and incrementally advanced to Grandmaster League). I know that the implementation of this might be really challenging, and it takes a lot of time for the AI to acquire its strategies (the so-called "agents" of AlphaStar are required to play thousands of games against each other and the best ones will then proceed to the next tiers). Now, since I am no expert on AI, and some of you might be more familiar with this topic, I only wanted to suggest that something like AlphaStar could be used, in future, for a very limited pool of nations - in this case the lucky nations. Some of you might have heard of AlphaStar, an AI developed by Google's DeepMind which recently managed to climb the ladder of StarCraft II (one of the most famous real-time strategy games) with highly remarkable results. This suggestion is clearly aiming at EUV and less EUIV since IV might be nearing its development cycle. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |