Presented at
VB2019,
Oct. 3, 2019, 9 a.m.
(60 minutes).
A welcome to the Threat Intelligence Practitioners' Summit from the Editor of Virus Bulletin, Martijn Grooten, followed by keynote address by F-Secure's Mika Stahlberg.
**Keynote: Fuelling AI with threat intelligence**
The cybersecurity industry has been data-driven - utilizing and sharing threat intelligence information - since its inception. Today's threat landscape contains larger volumes of more diverse and complex threats than ever before, and simply sharing file or URL samples is no longer enough to stay on top of the problem. The only way to adapt to the volumes of data we're receiving is to utilize machine learning techniques. Machine learning models are generally better when trained with large amounts of data, but they're only as good as the data they're trained against. When data is collected on a large scale it is not possible to properly vet everything by eye. Data collected in such a way may contain mislabelled samples, and possibly samples that were intentionally introduced in order to poison any model trained against them. When a model is trained with flawed data, it can behave unexpectedly, may exhibit biases, and most certainly won't provide the results we expect it to. How will working with large volumes of data and machine learning change the way cybersecurity practitioners collect, analyse, share, and use threat intelligence?
*(This presentation forms part of the Threat Intelligence Practitioners' Summit)*
Presenters:
-
Martijn Grooten
- Virus Bulletin
Mika Ståhlberg Mika Ståhlberg is the CTO of F-Secure. He joined F-Secure in 2004 and has served in various positions including Director of Security Research and VP of F-Secure Labs. He received a Master's degree in engineering from Helsinki University of Technology (Aalto University) as well as an Officer's degree from Finnish National Defence University.
-
Mika Ståhlberg
- F-Secure
Links:
Similar Presentations: