Agile. Efficient. Intuitive.

Introducing Starpoint™

The Next Generation of AI Deep Learning Cybersecurity

Quantum Star Technologies™ has developed a revolutionary and unique approach to securing the world’s data through deep learning. Starpoint’s focus is to enable customers to detect and classify threats proactively, decreasing time to detection (especially unknown zero-day threats), and increasing speed to quarantine. This Threat Hunting allows a business to run “pre-breach” analysis on their existing network and ensure they haven’t already been breached by a benign file.

What’s at Stake

93%

Of local networks can be penetrated by hackers

$3.92M

Avg cost of an enterprise data breach

1.5B

Number of IoT device breaches that took place Jan-June 2021

1,070%

Growth of ransomware between July 2020 and June of 2021

93%

Of local networks can be penetrated by hackers

$3.92M

Avg cost of an enterprise data breach

1.5B

Number of IoT device breaches that took place Jan-June 2021

56%

Growth of ransomware between July 2020 and June of 2021

Modern cyber attacks have proven that victims get ONE SHOT ONLY to identify and stop malware from reeking havoc on a systems database. 

And, in many cases, the threat already exists on the business’ network, just as a benign file, until it morphes into a serious cyber security threat.

Our Solution

Here’s how it works

Starpoint’s Core Technology, unpacks and inspects each file at its most granular level.

This allows us to identify any potentially harmful files with a <0.1%
False Positive Rate. Whether those files be new or pre-existing within a network.

Starpoint™ Threat Intelligence

Current cybersecurity solutions rely on static traditional means for detecting threats. Quantum Star Technologies™ has developed a revolutionary and unique approach to securing the world’s data through deep learning.

Starpoint™ enables customers to detect and categorize threats proactively by file type. This leads to quicker time to detection (especially unknown zero-day threats) by detecting the level of compromise and increasing speed to quarantine. 

Learn More

Starpoint™ uses “Pure Deep Learning” in its training and testing. This means there is no manual feature extraction or reverse engineering involved in the learning process, and we rely solely on deep learning models that use raw binary data.

This allows rapid adaptation in generating new models extraordinarily fast. Starpoint™ can detect malware that is written after deployment without any updates

Our process reorganizes the data so the deep learning networks can understand the data without human bias or error and because we use static analysis there is no need for file execution to detect maliciousness.

Due to these processes, Starpoint™ learns what malware is making it highly improbable for malicious code to circumvent detection.

Starpoint™ Detection

Using a proprietary and patented algorithm to detect characteristics of malicious code, Starpoint™ is essentially agnostic to data type, allowing it to be tailored to virtually any environment or system.  In the detection process, data is reorganized into a multidimensional coordinate system and then fed into an advanced neural network for evaluation.  These data points converge into information that Starpoint™ relays to the user as malicious or benign.  What traditionally took weeks or months, is reduced to minutes and seconds with Starpoint™.

And because it is so fast and lightweight, Starpoint™ can work double-time, inspecting your existing files for malware AND identifying any incoming threats.

Learn More

Starpoint™ uses processes to detect the maliciousness of a file by identifying and labeling different types of malware, including:

  • Ransomware
  • Spyware
  • Backdoors
  • Worms (Potentially Unwanted Application or PUA and Dropper under development)

This provides additional subjective threat information to the customer for an appropriate remediation response.

Starpoint™ Core

Starpoint™ is able to achieve a <0.1% False Positive Rate with NO prefiltering in place, while also requiring very low resource consumption to execute. Our proprietary Core Engine has low memory and compute consumption with average scan times of 23 milliseconds. The engine was designed to be deployed on any OS at any layer in the network – allowing for integration throughout a customer’s existing tech stack and offering protection independent of internet connectivity. This flexible deployment model allows for end-to-end protection for business and service platforms whether it is at the endpoint, hosting, or cloud layers.

Learn More

Instead of using industry standards and dependencies, we have created our own core engine that replicates frameworks and libraries in a more low-level code base. In doing so we have been able to obtain the following metrics:

  • Scan time per file total – 23 milliseconds (avg)
  • Scan time per file per model – <10 milliseconds
  • Efficiency of RAM consumption – 35 megabytes under full scan load
  • Footprint for Starpoint engine CLI (Command Line Interface) – 1.3 megabytes
  • Footprint for Starpoint™ engine sensor – 9.7 megabytes
  • Model(s) footprint total – <22 megabytes

Starpoint™ Data

All Starpoint’s services are powered by Starpoint™ Data, a data lake of malware threats integrated into a scalable deep learning pipeline. Starpoint™ Data allows for constant iteration, dynamically building and testing new models for deployment both in cloud and at edge.

We lead the industry in using multiple machine learning models in every evaluation. Our stacked deep learning AI is constantly learning and adapting to the changing threat landscape across all supported and potential future filetypes.

Learn More

Our SAAS API is scalable, highly accessible, standards compliant, and easy to integrate into existing pipelines.

Our Sensor takes advantage of Starpoint’s speed with no internet connection needed. This can be integrated into any pipeline that has constraints that Cloud options might have problems with.

Starpoint™ is the Best of breed for private and hybridized workflow.