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 | |
|---|---|---|---|---|
 |
6. oktober 2017 Kl. 12.12 55 min A146 |
Advancing Neuro-Fuzzy Algorithm for Automated Classification in Large-scale Forensic and Cybercrime Investigations Abstract: Big Data is a reality and Cyber Crime Investigators are confronted with the amount and complexity of seized digital data in criminal cases. Human experts are sitting in the Court of Law and making decisions with respect to found evidences that are being presented. Therefore, there is a strong need to bridge data processing and automated analysis for providing human-understandable representation of evidences. There is a history of successful applications of Machine Learning methods in Digital Forensics such that Artificial Neural Networks, Support Vector Machines and Bayes Network. However, the challenge is that such methods neither provide human-explainable models nor can work without prior knowledge required for inference and data representation. In this work Andrii focuses on Neuro-Fuzzy, a Hybrid Intelligence method that is capable of connecting two worlds: Computational Intelligence and Digital Forensics. CV: Andrii Shalaginov received his Master Degree from the Gjøvik University College in 2013 and also holds his degree from the National Technical University of Ukraine “Kyiv Polytechnic Institute” - Department of Computer Aided Design. Before studying at HiG he had an industry experience, including Samsung R&D center in Kiev. He joined NTNU Digital Group as a PhD student with the research topic related to application of soft computing in digital forensics. Andrii also has extensive knowledge in malware analysis and machine learning. |
Andrii Shalaginov | NISseminar |
 |
17. januar 2023 Kl. 12.15 105 min K102 |
Lecture 1 | Andrii Shalaginov | IMT4133 Data Science for Security and Forensics |
 |
31. januar 2023 Kl. 12.15 105 min K102 |
Lecture 2 | Andrii Shalaginov | IMT4133 Data Science for Security and Forensics |
 |
14. november 2014 Kl. 12.15 50 min K102 |
Scientific Seminar Soft Computing has been widely used in Information Security. Fuzzy Logic represents a particular interest since it is capable of not only classification, but also human-understandable representation of the information. Multiple studies show the applicability of the fuzzy logic in malware detection, yet the accuracy may not be sufficiently good classification accuracy due to the parameters of fuzzy sets. The talk will be about the adaptive Hybrid Neuro-Fuzzy methods that uses Self-Organizing Map to automatically extract parameters of fuzzy patches for the Mamdami-type classification rules. Moreover, the study of possible representation and visualization for reasoning will be given. |
Andrii Shalaginov | Scientific Seminar |