The success of kernel methods including support vector machines (SVMs) strongly depends on the design of appropriate kernels. While initially kernels were designed in order to han...
Semi-Supervised Support Vector Machines (S3VMs) typically directly estimate the label assignments for the unlabeled instances. This is often inefficient even with recent advances ...
Background: Document classification is a wide-spread problem with many applications, from organizing search engine snippets to spam filtering. We previously described Textpresso, ...
The proliferation of malware has presented a serious threat to the security of computer systems. Traditional signature-based antivirus systems fail to detect polymorphic and new, ...
Motivation: High-throughput technologies for rapid measurement of vast numbers of biological variables offer the potential for highly discriminatory diagnosis and prognosis; howev...