基于意图嵌入的社交机器人检测方法  

Intention embedding method based social bot detection

作　　者：牛红峰 李嘉伟 宋云鹏 蔡忠闽[1,2] NIU Hongfeng;LI Jiawei;SONG Yunpeng;CAI Zhongmin(Ministry of Education Key Lab for Intelligent Networks and Network Security,Xi’an Jiaotong University,Xi’an 710049,China;School of Automation Science and Engineering,Xi’an Jiaotong University,Xi’an 710049,China)

机构地区：[1]西安交通大学智能网络与网络安全教育部重点实验室,陕西西安710049 [2]西安交通大学自动化科学与工程学院,陕西西安710049

出　　处：《通信学报》2024年第11期194-205,共12页Journal on Communications

基　　金：国家自然科学基金资助项目(No.62102308)。

摘　　要：人工智能生成内容技术显著提升了社交机器人的伪装能力,给现有的机器人检测方法带来新的挑战。通过对社交平台用户意图进行建模,提出一种基于意图嵌入的社交机器人检测方法,从而避免直接在行为层面检测伪装能力大幅提升的机器人这一难题。实验结果表明,采用意图嵌入的检测模型相比未使用意图嵌入的模型,社交机器人检测准确率提高了5.58个百分点,并增强了对不同类型社交机器人的识别能力,验证了意图嵌入在提升人机检测任务性能中的有效性。Artificial intelligence generated content technology has significantly enhanced the disguise capabilities of so‐cial bots,presenting new challenges to existing bot detection methods.By modeling the intentions of social media users through intention representation,a intention embedding method based social bot detection was proposed,thereby avoid‐ing the difficulty of directly detecting bots with enhanced behavioral camouflage at the action level on social platforms.Experimental results show that the detection model using intention embedding improves the accuracy of social bot detec‐tion by 5.58 percentage points compared to models not utilizing intention embedding,and it enhances the recognition ca‐pability of specific types of social bots,verifying the effectiveness of intention embedding in improving the performance of human-bot detection tasks.

关 键 词：社交机器人检测 意图表征 意图嵌入 人工智能生成内容 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]