Silicon Valley-based cybersecurity startup PatternEx started back in 2013, but launched their Threat Prediction Platform in Februrary which is designed to detect anomalies and malicious data using supervised machine learning while also continuously incorporating input from human analysts. Cool Startup: PatternEx
The company was created as a result of two years of development by a team of AI and security experts and through building relationships with a few partnering companies in order to leverage their data for security testing. Based on testing using those real world data sets, PatternEx claims that the solution is able to detect 10 times more threats with five times fewer false positives compared with approaches based on Machine Learning-Anomaly Detection technology.
PatternEx’s solution, dubbed Active Contextual Modeling (ACM), parses huge amounts of data, generated by user activity, can search for anomalies using a recurrent neural network in combination with machine learning techniques. As a result, these models can learn from each other and achieve a network effect in detecting attack patterns.
Once anomalies are identified, the platform notifies a human analyst, presenting its findings report. The human analyst can determine whether a given behavior is malicious or not, which is reported back to the ACM. In turn, the ACM transforms raw data into behaviors, and the output is a predictive model that can assess in real time whether a given behavior is a rare event or not. The Threat Prediction Platform leverages these models to make real time predictions about specific threat vectors. The more attacks the system predicts, the more feedback it receives from the analysts, which in turn continually learns to detect existing and new attacks with improved accuracy and in real time in a closed-loop system.
With the emergence of AI-based malware solutions in the past year, PatternEx is definitely worth watching. So far as of July, they have raised a Series A investment of $7.8 million, led by Khosla Ventures.
- Company: PatternEx
- Started: 2013
- HQ: San Jose, CA
- VC Raised: $7.8M in 1 Round from 1 Investor
- Executives: Uday Veeramachaneni (CEO), George Jabrine (President) and Vamsi Korrapati (VP of Engineering)
- Employee Count: 11 – 50
- Product: Virutal Analyst Platform, or Human-Powered AI
- Customers: Mazda, Australian Geographic and Booktopia
- Big data platform designed for large data volumes and real time response
- Ensemble of algorithms designed to detect rare behaviors with the goal of identifying new attacks
- Mechanism to obtain feedback from security analysts and continuously update models with the provided feedback.
- Active learning feedback loop that continuously improves detection rates over time
- Threat intelligence reporting that can be shared among enterprises