BridgeHand2Vec Bridge Hand Representation

TL;DR

This paper introduces BridgeHand2Vec, a neural network-based method to embed bridge hands into a vector space, capturing hand strength and enabling various applications like reinforcement learning and bid classification.

Contribution

The novel BridgeHand2Vec approach creates interpretable vector representations of bridge hands, improving trick estimation and supporting AI applications in bridge.

Findings

Achieves state-of-the-art results on trick estimation

Provides meaningful hand similarity measures

Enables applications in reinforcement learning and bidding

Abstract

Contract bridge is a game characterized by incomplete information, posing an exciting challenge for artificial intelligence methods. This paper proposes the BridgeHand2Vec approach, which leverages a neural network to embed a bridge player's hand (consisting of 13 cards) into a vector space. The resulting representation reflects the strength of the hand in the game and enables interpretable distances to be determined between different hands. This representation is derived by training a neural network to estimate the number of tricks that a pair of players can take. In the remainder of this paper, we analyze the properties of the resulting vector space and provide examples of its application in reinforcement learning, and opening bid classification. Although this was not our main goal, the neural network used for the vectorization achieves SOTA results on the DDBP2 problem (estimating the number of tricks for two given hands).