I was greatly honored when I got an invitation from the Conference on Knowledge Discovery and Data Mining (KDD) to give a talk about data mining and cyber security.
Knowing me, you might be able to guess the topic I chose to present: Visual Analytics. I am focussing on not the visualization layer or the data layer, but on the analytics layer. In the presentation I am showing what we have been doing with data analytics and data mining in cyber security. The presentation starts out with an overview of what security is and what our data looks like. While I show a few examples for different areas in cyber security, I am mainly highlighting problems and challenges we have been facing within these areas with regards to analytics and data mining.
The presentation has 5 parts:
- Cyber Security - Lay of the Land: A quick introduction to the information / cyber security field.
- Data Mining in Security: For the data scientists out there, how does security data look like and what are some of the challenges you will face when dong data mining on security data (see slide below).
- Visual Analytics: This section discusses why is visual analytics a promising approach to the security data problem?
- Security Visualization: In three areas I am showing examples of visualization that we are using in the security field. I also outline the problems we are facing with the approaches.
- Challenges: This is a summary of some of the challenges we have in security data analytics. See below.
For each of the six areas in data mining, the following slide shows a couple of challenges that one will run into when trying to apply them to cyber security data:
At the end, I am presenting a number of challenges to the community; hard problems that we need help with to advance insights into cyber security of infrastructures and applications. The following slide summarizes the challenges I see in data mining for security:
Definitely not a complete list. Please comment and add other challenges! If you have any suggestions on solving the challenges, please contact me or comment on this post as well!