A A Quantum Ventura, Lockheed Martin, and Penn State InnovationQuantum Ventura’s CyberNeuro-RT (CNRT) technology offering has been developed in partnership with Lockheed Martin Co.’s MFC Division and Pennsylvania State University under partial funding from the U.S. Department of Energy.Cutting-Edge Unsupervised MLScalable Unsupervised Outlier Detection (SUOD)Large-scale heterogeneous outlier detection6 ML Algo EnsembleModel Approximation for Complex ModelsExecution Efficiency Improvement for Task Load Balancing in Distributed SystemVariational Autoencoder (VAE)Variational Autoencoder (VAE)Encoder-Decoder ArchitectureVariational => Highly Regularized EncoderETrained to Minimize Reconstruction Error of initial input and reconstructed outputVariational Autoencoder (VAE)75x Dataset Growth in Under 2 MonthsExisting Dataset Ingestion: Proprietary system enables ingestion of any existing network capture dataset with flexible support for any labelling systemFrom-the-wild Zero Day Sampling: System enables capturing and simulation of novel threats for additional data samplingData Generation via Simulation: ThreatATI database and proprietary ingestion system enable sampling and augmentation for cataloged threats from proprietary and public threat databasesProprietary Pipeline Adapts to Any DatasetThis is an illustration showing that the Proprietary Pipeline adapts to any datasetFollow Threats Home with Dark Web TrackingList of threats At-the-edge Neuromorphic Processing◯ Two offerings from the leading neuromorphic developers: Intel and Brainchip◯ Small form factor, magnitudes less power consumption than GPU◯ On-chip learning for deployment network specific attack detectionAn image showing Intel LoihiIntel LoihiA Brainchip AkidaBrainchip AkidaDashboards Minimizes Operator FatigueRobust, multi-faceted, user-friendly Cyber Analyst dashboard Operator fatigue allows cyber attacks to happenLarge numbers of false alarms cause real threats to be missedFalse alarms fatigue the cyber analyst further increasing risk of missed threats‍The Cyber Neuro-RT dashboard is designed to minimize all sources of analyst fatigue while presenting timely and meaningful data insightsAl based false alarms are minimized (trained for minimal false positive rate)n for cataloged threats from proprietary and public threat databasesPossible threats are ranked by importance and confidenceOnly the most relevant and likely alarms are actioned uponQuantum Ventura, Lockheed Martin, and Penn State Innovation
Quantum Ventura’s CyberNeuro-RT (CNRT) technology offering has been developed in partnership with Lockheed Martin Co.’s MFC Division and Pennsylvania State University under partial funding from the U.S. Department of Energy.
Cutting-Edge Unsupervised ML
Scalable Unsupervised Outlier Detection (SUOD)
- Large-scale heterogeneous outlier detection
- 6 ML Algo EnsembleModel Approximation for Complex Models
- Execution Efficiency Improvement for Task Load Balancing in Distributed System
- Variational Autoencoder (VAE)
Variational Autoencoder (VAE)
- Encoder-Decoder Architecture
- Variational => Highly Regularized Encoder
- ETrained to Minimize Reconstruction Error of initial input and reconstructed output
- Variational Autoencoder (VAE)
75x Dataset Growth in Under 2 Months
- Existing Dataset Ingestion: Proprietary system enables ingestion of any existing network capture dataset with flexible support for any labelling system
- From-the-wild Zero Day Sampling: System enables capturing and simulation of novel threats for additional data sampling
- Data Generation via Simulation: ThreatATI database and proprietary ingestion system enable sampling and augmentation for cataloged threats from proprietary and public threat databases
(20min delay)