Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

XQSV: A Structurally Variable Network to Imitate Human Play in Xiangqi

Chenliang Zhou

TL;DR

XQSV tackles the problem of reproducing human Xiangqi play by learning from human game data and dynamically adapting its network structure across Elo-based data subsets. It combines a Structurally Variable Recurrent Neural Network with four key innovations—Elo-range partitioning, sequential 1D input, imperfect memory capacity, and a locally illegal move filter—to achieve human-like move prediction. Empirically, XQSV attains about 40% overall accuracy, peaking within the trained Elo range, and a three-terminal Turing Test indicates outputs indistinguishable from real humans while outperforming a conventional Xiangqi app. The work also introduces relaxed evaluation metrics to accommodate non-deterministic human play and positions XQSV as a first benchmark for human-likeXiXiangi play, with potential applications in training and user experience.

Abstract

In this paper, we introduce an innovative deep learning architecture, termed Xiangqi Structurally Variable (XQSV), designed to emulate the behavioral patterns of human players in Xiangqi, or Chinese Chess. The unique attribute of XQSV is its capacity to alter its structural configuration dynamically, optimizing performance for the task based on the particular subset of data on which it is trained. We have incorporated several design improvements to significantly enhance the network's predictive accuracy, including a local illegal move filter, an Elo range partitioning, a sequential one-dimensional input, and a simulation of imperfect memory capacity. Empirical evaluations reveal that XQSV attains a predictive accuracy of approximately 40%, with its performance peaking within the trained Elo range. This indicates the model's success in mimicking the play behavior of individuals within that specific range. A three-terminal Turing Test was employed to demonstrate that the XQSV model imitates human behavior more accurately than conventional Xiangqi engines, rendering it indistinguishable from actual human opponents. Given the inherent nondeterminism in human gameplay, we propose two supplementary relaxed evaluation metrics. To our knowledge, XQSV represents the first model to mimic Xiangqi players.

XQSV: A Structurally Variable Network to Imitate Human Play in Xiangqi

TL;DR

XQSV tackles the problem of reproducing human Xiangqi play by learning from human game data and dynamically adapting its network structure across Elo-based data subsets. It combines a Structurally Variable Recurrent Neural Network with four key innovations—Elo-range partitioning, sequential 1D input, imperfect memory capacity, and a locally illegal move filter—to achieve human-like move prediction. Empirically, XQSV attains about 40% overall accuracy, peaking within the trained Elo range, and a three-terminal Turing Test indicates outputs indistinguishable from real humans while outperforming a conventional Xiangqi app. The work also introduces relaxed evaluation metrics to accommodate non-deterministic human play and positions XQSV as a first benchmark for human-likeXiXiangi play, with potential applications in training and user experience.

Abstract

In this paper, we introduce an innovative deep learning architecture, termed Xiangqi Structurally Variable (XQSV), designed to emulate the behavioral patterns of human players in Xiangqi, or Chinese Chess. The unique attribute of XQSV is its capacity to alter its structural configuration dynamically, optimizing performance for the task based on the particular subset of data on which it is trained. We have incorporated several design improvements to significantly enhance the network's predictive accuracy, including a local illegal move filter, an Elo range partitioning, a sequential one-dimensional input, and a simulation of imperfect memory capacity. Empirical evaluations reveal that XQSV attains a predictive accuracy of approximately 40%, with its performance peaking within the trained Elo range. This indicates the model's success in mimicking the play behavior of individuals within that specific range. A three-terminal Turing Test was employed to demonstrate that the XQSV model imitates human behavior more accurately than conventional Xiangqi engines, rendering it indistinguishable from actual human opponents. Given the inherent nondeterminism in human gameplay, we propose two supplementary relaxed evaluation metrics. To our knowledge, XQSV represents the first model to mimic Xiangqi players.
Paper Structure (18 sections, 1 equation, 1 figure, 4 tables)

This paper contains 18 sections, 1 equation, 1 figure, 4 tables.

Table of Contents

  1. Introduction and Background
  2. Board Game Engines
  3. Imitate Human in Board Games
  4. Xiangqi Engines
  5. Contributions
  6. XQSV Design Approach
  7. Data Preprocessing
  8. Elo Range Partitioning
  9. Sequential 1D Input
  10. Imperfect Memory Capacity
  11. Structurally Variable RNN
  12. Locally Illegal Move Filter
  13. Experiment and Evaluation
  14. Training XQSV on six Elo ranges
  15. Ablation Study
  16. ...and 3 more sections

Figures (1)

  • Figure 1: XQSV's prediction accuracy on different Elo ranges