Architecture Technology Corporation

Paul Nicotera

Judson Powers

Robert A Joyce

Scott Aloisio

Jafar Al-Gharaibeh

Colleen Kimball

Joseph Sirianni

William Estey

Matthew P Donovan

Eric R Chartier

Daniel McArdle

Dahyun Hollister

Katey Huddleston

Matthew A Stillerman

Kenneth McVearry

Ryan Marotz

Mandeep Singh

Daniel Mcardle

The methods and systems disclosed herein generally relate to automated execution and evaluation of computer network training exercises, such as in a virtual environment. A server executes a first attack action by a virtual attack machine against a virtual target machine based on a cyber-attack scenario, wherein the virtual target machine is configured to be controlled by the user computer. The server receives a user response to the first attack action, determines, using a decision tree, a first proposed attack action based on the user response, and executes an artificial intelligence model to determine a second proposed attack action based on the user response. The server selects a subsequent attack action from the first proposed attack action and the second proposed attack action and executes the subsequent attack action by the virtual attack machine against the virtual target machine.

System for cyber-attack simulation using artificial intelligence modeling

Disclosed herein are embodiments of systems, methods, and products comprise an analytic server, which provides a terrain segmentation and classification tool for synthetic aperture radar (SAR) imagery. The server accurately segments and classifies terrain types in SAR imagery and automatically adapts to new radar sensors data. The server receives a first SAR imagery and trains an autoencoder based on the first SAR imagery to generate learned representations of the first SAR imagery. The server trains a classifier based on labeled data of the first SAR imagery data to recognize terrain types from the learned representations of the first SAR imagery. The server receives a terrain query for a second SAR imagery. The server translates the second imagery data into the first imagery data and classifies the second SAR imagery terrain types using the classifier trained for the first SAR imagery. By reusing the original classifier, the server improves system efficiency.

Systems and methods of radar neural image analysis using nested autoencoding

Various embodiments described herein relate to a machine-learning based electronic media analysis software system. The system is configured to detect anomalous and predictive patterns associated with an event. The system is configured to use feature extraction techniques and semi-supervised machine-learning to detect the patterns associated with the event in the electronic media messages, which may indicate a synthetic driven behavior and conversation corresponding to the event.

Systems and methods for semantically detecting synthetic driven conversations in electronic media messages

Disclosed herein are embodiments of an optical character recognition pre-processing software system, which is integrated into a language translation system to provide automated cleaning and correction of noisy and degraded document images to enable seamless and efficient optical character recognition processing and machine translation of information within the document images.

Systems and methods for preprocessing document images

Disclosed herein are embodiments of systems, methods, and apparatus that execute classification techniques to enable high-quality analysis of ingest data by interpreting and categorizing disparate data points of the ingest data. The execution of the classification techniques leads to isolation of intrinsic properties of each data point to represent the essence of what the overall ingest data indicates. The classification techniques further enables classification of the ingest data, which is unencumbered by any ingest data format changes, such as ordering of data components, encoding, or properties associated with the ingest data that are likely to change without altering meaning conveyed by the ingest data.

Systems and methods for classification of data streams

Disclosed herein are embodiments of systems, methods, and products comprising an analytic server that automates training dataset generation for different application areas. The server may perform an automated, iterative refinement process to build a collection of dataset generator models over time. The server may receive a set of seed examples in a domain and generate candidate examples based on the features of the seed examples using data synthesis techniques. The server may execute a pre-trained label discriminator (LD) and domain discriminator (D2) on the candidate examples. The LD may identify and reject mislabeled data. The D2 may identify and reject out of domain data. The analytic server may regenerate new labeled data based on the feedback of the LD and D2. The analytic server may train a dataset generator by iteratively performing these steps for refinement until the regenerated candidate examples reach a pass rate threshold.

Systems and methods for machine learning dataset generation

In general, this disclosure describes methods and devices for analyzing source code to detect potential bugs in the code. Specifically, a device retrieves source code of an application. For each distinct execution of a plurality of executions of the application, the device initiates the respective execution at a particular starting point of the source code and inputs, into the source code, a unique set of inputs relative to any other execution. The device stores, into a path log, an indication of each line of source code and stores, into an output log, an indication of each output object encountered during the respective execution. Each output object includes a local variable dependent on the inputs. The device analyzes, using a machine learning model, the path and output logs to identify an abnormality indicative of a potential bug in the source code. The device outputs a graphical representation of the abnormality.

Automated software bug discovery and assessment

Systems and methods for document image processing using neural networks

The methods and systems disclosed herein generally relate to automated execution and evaluation of computer network training exercises, such as in a virtual environment. A server generates a training system having a virtual attack machine and a virtual target machine where the virtual target machine is operatively controlled by a trainee computer. The server then executes a simulated cyber-attack and monitors/collects actions and responses by the trainee. The server then executes an artificial intelligence model to evaluate the trainee's action and to identify a subsequent simulated cyber-attack (e.g., a next step to the simulated cyber-attack). The server may then train the artificial intelligence model using various machine-learning techniques using the collected data during the exercise.

Artificial intelligence modeling for cyber-attack simulation protocols

Systems and methods for used learned representations to determine terrain type

An example technique includes initializing, by an obfuscation computing system, communications with nodes in a distributed computing platform. The nodes include compute nodes that provide resources in the distributed computing platform and a controller node that performs resource management of the resources. The obfuscation computing system serves as an intermediary between the controller node and the compute nodes. The technique further includes outputting an interactive user interface (UI) providing a selection between a first privilege level and a second privilege level, and performing one of: based on the selection being for the first privilege level, a first obfuscation mechanism for the distributed computing platform to obfuscate digital traffic between a user computing system and the nodes, or based on the selection being for the second privilege level, a second obfuscation mechanism for the distributed computing platform to obfuscate digital traffic between the user computing system and the nodes.

Obfuscation for high-performance computing systems

Techniques for performing data analytics using anomaly detection systems and methods are disclosed. The anomaly detection system provides an incident response and monitoring solution, built for distributed processing, that streamlines cyber defense by unifying datasets, via a data translator, from sensors and tools into a uniform schema to provide real-time anomaly detection, via an anomaly detection system that may prevent malware from establishing a foothold on the network. The anomaly detection system may allow for the scalability to provide large-scale data aggregation and anomaly detection without compromising performance. The anomaly detection system may use a distributed architecture to support advanced cyber threat detection across large datasets in real-time for monitoring and rapid incident response. The anomaly detection system may leverage open protocols and interfaces to promote third-party support for development and interoperability.

Scalable incident-response and forensics toolkit

A computer-implemented method includes creating, by a computing device, an abstract syntax tree based on a source code file of a software application, the source code file including source code defining operations of the software application. The method also includes traversing, by the computing device, the abstract syntax tree. The method further includes identifying, by the computing device and based on the traversing of the abstract syntax tree, one or more code violations present in the source code. The method also includes generating, by the computing device, at least one refactoring option for the one or more code violations, each refactoring option of the at least one refactoring option representing a change to the source code file that is configured to remediate the associated code violation.

Growing community of inventors

Ithaca, NY, United States of America

Paul Nicotera