Relational AI Plays Bridge
Understanding the complexity of each person in relation to the world around them.
Get an exclusive first look at the Relational AI BRIDGE Proof-of-Concept
Why the game of Bridge?
No AI has proven to learn and play bridge well because it requires understanding both logical game play and communications between players to predict if players are telling the truth or playing a trick. The human relational aspects of emotions and communications makes it a much more complex adaptive system than Go or chess.
Understanding Deeper Human Reasoning
While strict rules govern what can be said and what actions can be taken on a player’s turn, each move carries with it deeper meanings that need to be understood by all involved. A player must also be able to justify their reasoning behind a particular move (no blackbox AI’s allowed here).
Complex Decision Making
Our bridge player learns from within game actions and human communications combined with external memory of previous game play and expert commentary from the internet. It uses natural language processing and reinforcement learning in our Geometric Associative Memory (GAM) to form a “hyperbolic lens” representing how each player perceives the world. It uses these lenses to interpret the moves and decide what actions to take in a complex decision-making process.
Get an exclusive first look at the Relational AI learning BRIDGE