本篇論文主要的目的是要解決電腦網路遊戲中的外掛非法程式問題，外掛非法程式問題一直是電腦網路遊戲營運商的一大困擾。所謂的外掛非法程式便是機器人程式(或稱代理人程式)，玩家角色可以透過外掛程式來自動的執行遊戲中的行為，玩家不需親自花時間來訓練角色的成長，進而從中獲得比其他玩家較多的不公平利益。外掛程式的氾濫導致玩家對營運商失去信心而離開遊戲，外掛程式所造成玩家的流失對營運商來說是營利及聲譽的雙重損失。
針對外掛程式問題，本篇論文提出一完整流程來解決外掛程式問題：1、「外掛玩家篩檢」，使用資料探勘技術來搜尋可能使用外掛程式的玩家，並將可能為外掛玩家的人數從所有線上玩家濃縮至少數玩家；2、「自動化外掛玩家辨別」，在這裡我們採用CAPTCHA(Completely Automated Public Turing Test to Tell Computers and Humans Apart)機制由系統自動的對可能為外掛程式之玩家進行是否為外掛程式的偵測；3、「外掛玩家追蹤」，透過遊戲管理者的追蹤可以避免系統誤判真實玩家為外掛程式的情況；4、「外掛玩家處罰」，若玩家角色確實為外掛程式在執行其行動則必需接受處罰。

Computer online games are popular and have good prospects of investment profit to on-line game producers in recent years. Since playing on-line games are very competitive and time-consuming, many players try to use illegal software agent (we call it BOT) to play for them. Thus, BOTs can automatically train players’ game roles and get advanced ranks for their game roles.
The goal of our research is to detect the Bots used in games and provide a standard management process to handle BOT-detected situation. The research constructs a complete process which includes historical data mining/analysis, BOT identification, BOT confirmation and player punishment mechanism. In historical data mining/analysis, we use the decision tree method to make our BOT-detection rule that could filter out possible BOT players. In BOT identification stage, we implement CAPTCHA mechanism to inspect whether an online game player is a BOT or not. In BOT confirmation stage, field-side game manager will trace and check the potential Bot player who is identified during the BOT identification stage to prevent false-positive situation. Once a BOT players are confirmed, player punishment mechanism will be activated.

[1] 外掛的歷史資訊 http://game.tzinfo.net/news_info.asp?id=10194
[2] Samuel Johnson, BOTKNOWLEDGE http://www.botknowledge.com/
[3] School of Computer Science, Carnegie Mellon, HIPs Project http://www.aladdin.cs.cmu.edu/hips/
[4] A. M. Turing ,”Computing Machinery and Intelligence”. Mind 49: 433-460,1950
[5] School of Computer Science, Carnegie Mellon, CAPTCHA Project http://www.captcha.net/
[6] Eric Wan, Xin Xu and Zhengqian Zhou.”2006 Online Game Report”, Pacific Epoch Red Innovation Report Series.
[7] Pinkas, B. and Sander, T. “Securing passwords against dictionary attacks”. Proceedings of the ACM Computer and Security Conference 2002
[8] Rachna Dhamija and J.D. Tygar. “Phish and HIPs: Human Interactive Proofs to Detect Phishing Attacks”. Second International Workshop (HIPs 2005)
[9] Amalia Rusu and Venu Govindaraju. “Handwritten CAPTCHA: Using the difference in the abilities of humans and machines in reading handwritten words”. Proceedings of the 9th Int’l Workshop on Frontiers in Handwriting Recognition (IWFHR-9 2004)
[10] Allison L. Coates, Henry S. Baird and Richard J. Fateman. “Pessimal Print: A Reverse Turing Test”. ICDAR 2001
[11] JOCR Project , Joerg Schulenburg http://jocr.sourceforge.net/download.html
[12] Yong Rui and Zicheg Liu. “ARTiFACIAL: Automated Reverse Turing Test Using FACIAL Features”, Proceedings of ACM Multimedia 2003
[13]Tsz-Yan Chan. “Using a Text-to-Speech Synthesizer to generate a reverse Turing Test”. Proceedings of the 15th IEEE International Conference on Tools with Artificial Intelligence (ICTAI’03)
[14]Leo Breiman [et al.]. “Classification and Regression Tree”. Wadsworth International Group, 1984.
[15]Philippe Golle and Nicolas Ducheneaut. “Preventing Bots from Playing Online Games”. ACM Computers in Entertainment,Vol.3 No.3 July 2005
[16]Wenfeng Yang and Xing Li.”Chinese Keyword Extraction Based on Max-duplicated Strings of the Documents”.ACM Research and Development in Information Retrieval, 2002.
[17]David Goldberg, Christopher Malon and Marshall Bern. “A Global Approach to Automatic Solution of Jigsaw Puzzles”. ACM Proceedings of the 1st conference on Computing frontiers. April 2004.
[18]Computer Vision and Artificial Intelligence http://www.acm.org/crossroads/xrds3-1/vision.html