Most scientists agree that no one really knows how the most advanced algorithms do what they do, nor how well they are doing it. That could be a problem. Advances in synthetic data generation technologies can help. These algorithms generate data with a known ground truth, sufficient volumes and with statistically relevant true and false positives (TP, FP) and true and false negatives (TN, FN) for the nature of the test. AI algorithms can now be measured for precision, c, as the fraction of the predicted matches that are true positive matches, or c = TP/(TP + FP).
In recent news, Pitney Bowes and Groupe M6 experienced ransomware attacks which limited customer access to company services and led to the encryption of information on private networks and systems belonging to the companies. Furthermore, email servers and phone lines also went down due to the attacks, and while no customer data was lost or stolen, shows how much of a threat these ransomware attacks can pose on the privacy of companies and their customers.
Ransomware attacks, while hard to detect and fight off, are able to be defeated with time and effort. However, if it takes too much time to defeat said attacks, valuable data could be breached or stolen and many will be put at risk. If the risk is too much, companies forego hopes of fighting off the attacks themselves and end up paying high extortion fees to minimize damage. However what happens when attackers strike again? Will the companies be prepared to fend it off the next time, or will be they be seen as an easy target because they gave in?
One thing is for sure; just as we continue to make strides in the cyber security industry, criminals continue to get more and more advanced with their own cyber attack tactics.
With the recent Equifax breach coming back into the limelight due to the cancellation of the $125 check the FTC promised to those impacted by the breach, we want to take a look at possible prevention for the breach in the first place, or at least ways that the damage could have been minimized.
Earlier this month, the heart of Manhattan was struck with a major power outage estimated to have impacted up to 72,000 Con Edison customers. While dangerous and definitely hard to look on the bright side of things, there are reports that do bring good news concerning the power outage. NBC reports that terrorism and cyber-attacks were ruled out following an investigation ordered by Mayor Bill De Blasio. So what could have caused this major blackout?
Our engineers have recently created a script that will inject synthetic data that simulates ADAMS data into a file format that can be consumed by commercial network traffic generators. ADAMS data is simulated data for insider threat detection systems based on anomalies in massive data-sets. Data domains include Logon, Device, HTTP, Email, File, Print, LDAP, Organization Directory, Decoy files, and Psychometric files. Why all of the excitement? The current state of the art network traffic generation tools are using very simplistic content that are not designed for the system under test. Once this integration is complete, cyber security testing can be taken to a while new level where sophisticated threat patterns are interwoven into data and consumed by the network. This will enable sophisticated testing of the network's intrusion detection and measurement of true and false positive errors, so these systems can be optimized for cost and risk performance. This alone is a huge leap in the cyber security industry, and we will only continue to move forward with our advancements in the world of technology.
High interest helping to implement Cyber Behavioral Tools was expressed by many potential clients, including the Cyber Innovation Manager from one of the world's largest banks, a Divisional Chief Information Security Officer for one of the biggest US Federal Systems Integrator's and one of the largest Cyber Independent Testing Laboratories. During the demonstrations large amounts of internally consistent data was generated for all desired behaviors. Data was generated over any time-frame to output: