Dark Reading is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them.Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Threat Intelligence

1/11/2018
10:30 AM
Raffael Marty
Raffael Marty
Commentary
Connect Directly
Twitter
RSS
E-Mail vvv

AI in Cybersecurity: Where We Stand & Where We Need to Go

How security practitioners can incorporate expert knowledge into machine learning algorithms that reveal security insights, safeguard data, and keep attackers out.

This visualization of 100GB of network traffic over a period of one week is used to determine 'normal' behavior and identify potential  anomalies by experts.
This visualization of 100GB of network traffic over a period of one week is used to determine 'normal' behavior and identify potential anomalies by experts.

Comment  | 
Print  | 
Comments
Newest First  |  Oldest First  |  Threaded View
cknisley44
100%
0%
cknisley44,
User Rank: Apprentice
1/12/2018 | 4:37:14 PM
Bayesian Belief Networks for Cybersecurity
Really nice article highlighting the exact things we at Haystax Technology talk about with the use of AI in cybersecurity. It seems every vendor has launched a new version of their product in the past year with the tagline of "New and improved with AI". 

There are great places where ML is perfect for cybersecurity, but only as a part of the solution. The heart of our Constellation Analytics Platform is a Bayesian Inference Network built over years of research to reason like a team of cyber experts across weak inputs from disparate sensor systems. Some of those sensors need ML techniques to find the real signal, some don't - but they all come together in a BayesNet that prioritizes the threat and ultimately the risk. 

We believe this is the only viable approach today given what you highlighted of lack of training data, and even if you had some training data, sensor outputs change frequently as the environment changes. Only with additional context can you uncover those malicious events that represent the highest risks. 
cchio
100%
0%
cchio,
User Rank: Apprentice
1/11/2018 | 11:38:23 PM
Great post!
Enjoyed your post. Agree with many of the points you made. Think you will resonate with a lot we have to say in the book we are publishing next month - https://www.amazon.com/Machine-Learning-Security-Protecting-Algorithms/dp/1491979909 Happy to chat more!
44% of Security Threats Start in the Cloud
Kelly Sheridan, Staff Editor, Dark Reading,  2/19/2020
Zero-Factor Authentication: Owning Our Data
Nick Selby, Chief Security Officer at Paxos Trust Company,  2/19/2020
Register for Dark Reading Newsletters
White Papers
Video
Cartoon
Current Issue
6 Emerging Cyber Threats That Enterprises Face in 2020
This Tech Digest gives an in-depth look at six emerging cyber threats that enterprises could face in 2020. Download your copy today!
Flash Poll
How Enterprises Are Developing and Maintaining Secure Applications
How Enterprises Are Developing and Maintaining Secure Applications
The concept of application security is well known, but application security testing and remediation processes remain unbalanced. Most organizations are confident in their approach to AppSec, although others seem to have no approach at all. Read this report to find out more.
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
CVE-2020-9405
PUBLISHED: 2020-02-26
IBL Online Weather before 4.3.5a allows unauthenticated reflected XSS via the redirect page.
CVE-2020-9406
PUBLISHED: 2020-02-26
IBL Online Weather before 4.3.5a allows unauthenticated eval injection via the queryBCP method of the Auxiliary Service.
CVE-2020-9407
PUBLISHED: 2020-02-26
IBL Online Weather before 4.3.5a allows attackers to obtain sensitive information by reading the IWEBSERVICE_JSONRPC_COOKIE cookie.
CVE-2020-9398
PUBLISHED: 2020-02-25
ISPConfig before 3.1.15p3, when the undocumented reverse_proxy_panel_allowed=sites option is manually enabled, allows SQL Injection.
CVE-2015-5201
PUBLISHED: 2020-02-25
VDSM and libvirt in Red Hat Enterprise Virtualization Hypervisor (aka RHEV-H) 7-7.x before 7-7.2-20151119.0 and 6-6.x before 6-6.7-20151117.0 as packaged in Red Hat Enterprise Virtualization before 3.5.6 when VSDM is run with -spice disable-ticketing and a VM is suspended and then restored, allows r...