🔥 The LAST two days to subscribe to TheSequence Edge with a unique 60% discount. Subscribe and share with your colleagues! Thank you for your support.🔥

SUBSCRIBE FOR $20/year

💥 What’s New in AI: DeepMind’s New Super Model that can Master Perfect and Imperfect Information Games 

Gaming has been one of the areas at the center of the deep learning renaissance of the last few years. This is not particularly surprising considering that games are environments relatively easy to reproduce in order to mimic scenarios of the real world. Despite the advancements, ML for gaming has been specialized in either perfect or imperfect information games but never both at the same time. Models that mastered chess and Go struggle with imperfect games such as Poker. Even models such as AlphaZero that learned to play multiple games simultaneously were constrained to perfect information environments. The reason for this is based on the intrinsic dynamics of both types of game environments. Perfect information games such as Chess and Poker are a good fit for ML techniques like self-play learning and tree search for the problem space, while imperfect information games like Poker rely on game-reasoning techniques. A few weeks ago, DeepMind challenged that conventional approach by proposing a neural network that is able to master both perfect and imperfect information games →let’s dive in to learn about the predecessors and how the PoG Algorithm works