Countering Good Word Attacks on Statistical Spam Filters with Instance Differentiation and Multiple Instance Learning
Author | : Yan Zhou |
Publisher | : |
Total Pages | : |
Release | : 2008 |
ISBN-10 | : 9537619036 |
ISBN-13 | : 9789537619039 |
Rating | : 4/5 (36 Downloads) |
Download or read book Countering Good Word Attacks on Statistical Spam Filters with Instance Differentiation and Multiple Instance Learning written by Yan Zhou and published by . This book was released on 2008 with total page pages. Available in PDF, EPUB and Kindle. Book excerpt: A multiple instance learning counterattack strategy for combating adversarial good word attacks on statistical spam filters has been presented. In the proposed strategy, e-mails are treated as multiple instance bags and a logistic model at the instance level is learned indirectly by maximizing the bag-level binomial log-likelihood function. The proposed counterattack strategy has been demonstrated on good word attacks of varying strength and has been shown to be effective. Additionally, we have confirmed earlier reports that.