兵棋仿真环境下基于W-GNN的小样本意图识别模型

宣自卓; 张永亮; 周献中; 孙宇祥

doi:10.12382/bgxb.2025.0363

您当前的位置：

首页 >

文章列表页 >

兵棋仿真环境下基于W-GNN的小样本意图识别模型

更新时间：2026-04-24

- 兵棋仿真环境下基于W-GNN的小样本意图识别模型
- A Few-shot Intention Recognition Model Based on W-GNN in a Wargame Simulation Environment
- 兵工学报 2026年47卷第3期页码：250363
- 作者机构：
  
  南京大学工程管理学院，江苏南京 210008
  陆军工程大学江苏南京 210001
  南京大学机器人与自动化学院江苏苏州 215163
- 作者简介：
  
  通信作者邮箱：sunyuxiang@nju.edu.cn
- 基金信息：
  
  国家自然青年基金项目(62306135);教育部青年基金项目(23YJC630156);江苏省青年基金项目(BK20230783);广东省重点实验室项目(GPKLPSNS-2024-KF-01)
- DOI：10.12382/bgxb.2025.0363
  中图分类号： TP391.9
- 收稿：2025-05-13，
  
  网络首发：2026-01-27，
  
  纸质出版：2026-03
- 稿件说明：
移动端阅览
宣自卓, 张永亮, 周献中, 等. 兵棋仿真环境下基于W-GNN的小样本意图识别模型[J]. 兵工学报, 2026,47(3):250363.

XUAN Zizhuo, ZHANG Yongliang, ZHOU Xianzhong, et al. A Few-shot Intention Recognition Model Based on W-GNN in a Wargame Simulation Environment[J]. Acta Armamentarii, 2026, 47(3): 250363.
宣自卓, 张永亮, 周献中, 等. 兵棋仿真环境下基于W-GNN的小样本意图识别模型[J]. 兵工学报, 2026,47(3):250363. DOI： 10.12382/bgxb.2025.0363.

XUAN Zizhuo, ZHANG Yongliang, ZHOU Xianzhong, et al. A Few-shot Intention Recognition Model Based on W-GNN in a Wargame Simulation Environment[J]. Acta Armamentarii, 2026, 47(3): 250363. DOI： 10.12382/bgxb.2025.0363.

摘要

传统意图识别模型通常依赖大规模战斗数据进行建模与训练，但在电子对抗与隐身技术不断发展的背景下，情报信息获取受限，使基于大样本的数据驱动方法面临适应性不足的问题。针对上述挑战，引入小样本学习思想至意图识别研究中，将任务建模为监督式消息传递过程，构建一种融合双向长短期记忆（Bidirectional Long Short-term Memory，BiLSTM）网络与部分可观测图模型的端到端深度学习架构。利用BiLSTM网络从有限兵棋态势信息中提取关键时序特征，刻画动态演化规律，并在此基础上构建加权图结构，通过图卷积实现节点特征更新与关系建模，最终完成意图判别。基于兵棋推演平台，在不同情报完备度条件下开展在线意图识别实验，并对比分析特征提取器配置对识别性能的影响。实验结果表明，在数据稀缺场景下，该模型仍具备良好的识别精度与鲁棒性，整体性能优于多种典型小样本学习模型，体现了其在智能指挥与决策支持中的应用潜力。

Abstract

Traditional intention recognition methods typically rely on large-scale combat data for modeling and training. However

with the continuous development of electronic countermeasures and stealth technologies

the acquisition of intelligence information has been increasingly limited

making the data-driven methods based on large samples difficult to adapt to modern operational environment. To address this challenge

a few-shot learning is introduced into the intention recognition task

which is formulated as a supervised message-passing process. A end-to-end deep learning architecture that integrates a bidirectional long short-term memory（BiLSTM）network with a partially observable graph model is developed. The BiLSTM network is employed to extract the essential temporal features from limited wargaming situation information to capture the dynamic evolution patterns

while suppressing noise and redundant information. A weighted graph structure is constructed based on the extracted features and the observed intention labels

and a graph convolution is applied to model the relational dependencies and iteratively update the node representations

ultimately achieving the intention predictions. Online intention recognition experimentsonline intention recognition experiments are conducted on a wargaming simulation platform under different levels of intelligence completeness

and the influence of feature extractor configuration on recognition performance is analyzed. Experimental results demonstrate that the proposed model maintains robust and accurate intention recognition performance under data-scarce conditions

outperforming several representative few-shot learning models and highlighting its potential applications in intelligent command and decision-support.

关键词

Keywords

references

CHEN L,LIANG X X,FENG Y H,et al. Online intention recognition with incomplete information based on a weighted contrastive predictive coding model in wargame[J]. IEEE Transactions on Neural Networks and Learning Systems,2023,34(10):7515-7528.

傅妍芳,雷凯麟,魏佳宁,等.基于演员-评论家框架的层次化多智能体协同决策方法[J].兵工学报,2024,45(10):3385-3396.

FU Y F, LEI K L, WEI J N, et al. Hierarchical multi-agent cooperative decision-making method based on the actor-critic framework[J]. Acta Armamentarii,2024,45(10):3385-3396. (in Chinese)

张大永,杨镜宇.面向兵棋推演复盘分析的机器学习数据集构建[J].系统仿真学报,2024,36(3):608.

ZHANG D Y, YANG J Y. Construction of a machine learning dataset for wargame replay analysis[J]. Journal of System Simulation,2024,36(3):608. (in Chinese)

GRAVES A,WAYNE G,DANIHELKA I. Neural turing machines:arXiv:1410.5401[R/OL]. Ithaca, NY, US: Cornell University, 2014(2014-12-10). https:∥arxiv. org/abs/1410.5401.

SANTORO A,BARTUNOV S,BOTVINICK M,et al. Meta-learning with memory-augmented neural networks[C]∥Proceedings of the 33rd International Conference on Machine Learning. New York, NY,US:PMLR,2016:1842-1850.

陈露,尚家兴,刘大江,等.基于异构图神经网络的可解释兵棋态势预测方法[J].自动化学报,2025,51(6):1248-1260.

CHEN L, SHANG J X, LIU D J, et al. Interpretable wargame situation prediction method based on heterogeneous graph neural networks[J].Acta Automatica Sinica,2025,51(6):1248-1260. (in Chinese)

MUNKHDALAI T,YU H. Meta networks[C]∥Proceedings of the 34th International Conference on Machine Learning. Sydney, Australia:PMLR,2017:2554-2563.

FINN C,ABBEEL P,LEVINE S. Model-agnostic meta-learning for fast adaptation of deep networks[C]∥Proceedings of the 34th International Conference on Machine Learning. Sydney,Australia:PMLR,2017:1126-1135.

SUN Q,LIU Y,CHUA T S,et al. Meta-transfer learning for few-shot learning[C]∥Proceedings of the 32nd IEEE/CVF Conference on Computer Vision and Pattern Recognition. Long Beach,CA,US:IEEE,2019:403-412.

NI R,GOLDBLUM M,SHARAF A,et al. Data augmentation for meta-learning[C]∥Proceedings of the 38th International Conference on Machine Learning. Virtual Conference: PMLR, 2021:8152-8161.

BRUNA J,ZAREMBA W,SZLAM A,et al. Spectral networks and locally connected networks on graphs:arXiv:1312.6203[R/OL]. Ithaca,NY, US: Cornell University,2014 (2014-05-21). https:∥arxiv. org/abs/1312.6203.

HENAFF M,BRUNA J,LECUN Y. Deep convolutional networks on graph-structured data: arXiv:1506.05163[R/OL]. Ithaca, NY,US: Cornell University, 2015 (2015-6-16). https:∥arxiv. org/abs/1506.05163.

DEFFERRARD M, BRESSON X, VANDERGHEYNST P. Convolutional neural networks on graphs with fast localized spectral filtering[C]∥Proceedings of the 30th International Conference on Neural Information Processing Systems. Barcelona, Spain:Curran Associates,Inc. ,2016:3844-3852.

KIPF T N, WELLING M. Semi-supervised classification with graph convolutional networks: arXiv: 1609.02907[R/OL]. Ithaca,NY,US:Cornell University,2017(2017-02-22). https:∥arxiv. org/abs/1609.02907.

YOU Y N, CHEN T L, WANG Z Y, et al. When does self-supervision help graph convolutional networks?[C]∥Proceedings of the 37th International Conference on Machine Learning. Virtual Conference:PMLR,2020:10871-10880.

BO D Y, WANG X, SHI C, et al. Beyond low-frequency information in graph convolutional networks[C]∥Proceedings of the 35th AAAI Conference on Artificial Intelligence. Virtual Conference:AAAI,2021:3950-3957.

JIN W,LIU X R, MA Y, et al. Feature overcorrelation in deep graph neural networks:a new perspective[C]∥Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York, NY, US: Association for Computing Machinery,2022:709-719.

CHEN C, LI K L, WEI W, et al. Hierarchical graph neural networks for few-shot learning[J]. IEEE Transactions on Circuits and Systems for Video Technology,2022,32(1):240-252.

FILIPIAK D,FENSEL A,FILIPOWSKA A. KGTN-ens:few-shot image classification with knowledge graph ensembles[J]. Applied Intelligence,2024,54(2):1893-1908.

GANESAN P,JAGATHEESAPERUMAL S K,HASSAN M M,et al. Few-shot image classification using graph neural network with fine-grained feature descriptors[J]. Neurocomputing, 2024, 610:128448.

FENG M, MEUNIER J. A lightweight graph neural network algorithm for action recognition based on self-distillation[J]. Algorithms,2023,16(12):552.

TORRES L, ARRAIS J P, RIBEIRO B. Few-shot learning via graph embeddings with convolutional networks for low-data molecular property prediction[J]. Neural Computing and Applications,2023,35(18):13167-13185.

GRAVES A. Long short-term memory[M]∥GRAVES A. Supervised Sequence Labelling with Recurrent Neural Networks. Berlin,Heidelberg,Germany:Springer,2012:37-45.

GILMER J, SCHOENHOLZ S S, RILEY P F, et al. Neural message passing for quantum chemistry[C]∥Proceedings of the 34th International Conference on Machine Learning. Sydney, Australia:PMLR,2017:1263-1272.

SONG Y,WANG T,CAI P,et al. A comprehensive survey of few-shot learning:evolution,applications,challenges,and opportunities[J]. ACM Computing Surveys,2023,55(13s):271.

浏览量

下载量

CNKI被引量

文章被引用时，请邮件提醒。

提交

工具集

关联资源

暂无数据