Dette systemet er ikke lenger i aktivt bruk og har blitt erstattet av Panopto i løpet av sommeren 2023. Les mer her.
| Dato | Tittel | Foreleser | Emne | |
|---|---|---|---|---|
 |
19. mai 2017 Kl. 12.12 55 min K105 |
A novel Binarization Scheme for Real-valued Biometric Feature Biometric binarization is the feature-type transformation that converts a specific feature representation into a binary representation. It is a fundamental issue to transform the real-valued feature vectors to the binary vectors in biometric template protection schemes. The transformed binary vectors should be high for both discriminability and privacy protection when they are employed as the input data for biometric cryptosystems. We present a novel binarization scheme based on random projection and random Support Vector Machine (SVM) to further enhance the security and privacy of biometric binary vectors. The proposed scheme can generate a binary vector of any given length as an ideal input for biometric cryptosystems. In addition, the proposed scheme is independent of the biometric feature data distribution. Several comparative experiments are conducted on multiple biometric databases to show the feasibility and efficiency of the proposed scheme. |
Jialiang Peng | NISseminar |
 |
15. september 2017 Kl. 12.12 55 min A146 |
Big Data Analytics: Topic Modeling for Digital Forensics Investigations and Cyber Threat Intelligence “Big Data Analytics” has become a high priority topic in Cyber Research and in the field of Cyber Security, Big Data represents a very serious problem. In the domain of Digital Forensics Investigations (DFI), the sheer volume of data to be analyzed impedes police operations that require timely reporting of DFI results to support active criminal investigations in the field. In the domain of Cyber Threat Intelligence (CTI), a rapid assessment of the available threat data is required to enable dissemination of actionable intelligence in a timely manner. Topic Modeling is an unsupervised machine learning method for analyzing large bodies of text data and producing estimates of the topics under discussion in them. To gain some insight into how it works, we reviewed some of the underlying principles of Topic Modeling. Then, I presented experimental results that show how Topic Modeling would work in the specific domains of DFI (using the Enron data set) and CTI (using posts scraped from an online hacker forum). |
Carl Stuart Leichter | NISseminar |