Author(s): Alireza Saberi, Yangchun Fu, Zhiqiang Lin

Download: Paper (PDF)

Date: 22 Feb 2014

Document Type: Briefing Papers

Additional Documents: Slides

Associated Event: NDSS Symposium 2014


Recent advances show that we can reuse the legacy binary code to bridge the semantic gap for VMI. However, existing solutions often have high overhead. This paper presents Hybrid-Bridge, a new system that uses decoupled execution and training memorization to bridge the semantic-gap, resulting in one order of magnitude faster according to our experimental results.