Google Inc.

Mandiant Corporation

Scott Eric Coull

Jeffrey Thomas Johns

David Krisiloff

Brian Sanford Jones

Giorgio Severi

A cyber-security analysis method uses machine learning (ML) technology to classify cyber-threat indicators, for example, as malicious or benign, by generating a threat score. The method includes receiving, at a compute device, a data set including cyber-threat indicators and verdicts serving as votes from each source in the set of sources. Each of the votes is associated with one of the cyber-threat indicators. An ML model is trained based on at least one of agreements among the sets of votes, and disagreements among the sets of votes to produce a trained ML model. In response to receiving a new cyber-threat indicator, votes are identified for each source from a subset of the sources, to define a second set of votes. The cyber-threat score is generated for the new cyber-threat indicator based on the trained ML model and the second set of votes.

Cyber-threat score generation using machine learning and reflecting quality of sources

A method for performing cyber-security analysis includes generating a semantic graph in which each object is represented as a node, and each event associated with an object is represented as an edge. A cyber-threat related alert, with an associated alert type, is received from a source. A first object from the plurality of objects is modified based on the alert. A plurality of threat scores, each associated with an object, are calculated, substantially concurrently, based on the alert type. Subsequently, a plurality of modified threat scores are determined for each object, based on: (1) the threat score for that object, (2) a connectivity of that object to each of the remaining objects within the semantic graph; and (3) the threat score for each remaining object from the plurality of objects. A subgraph of the semantic graph is identified based on normalized versions of the modified threat scores.

System and method for cybersecurity threat monitoring using dynamically-updated semantic graphs

A system for detecting whether a file including content s associated with a cyber-attack is described. The content may include an executable file for example. The system includes an intelligence-driven analysis subsystem and a computation analysis subsystem. The intelligence-driven analysis subsystem is configured to (i) receive the file, (ii) inspect and compute features of the file for indicators associated with a cyber-attack, and (iii) produce a first output representing the detected indicators. The computational analysis subsystem includes an artificial neural network to (i) receive a network input being a first representation of at least one section of binary code from the file as input, and (ii) process the first representation of the section to produce a second output. The first output and the second output are used in determination a classification assigned to the file.

System and method for analyzing binary code for malware classification using artificial neural network techniques

A method includes training a first machine learning model with a first dataset, to produce a first trained machine learning model to infer cybersecurity-oriented file properties and/or detect cybersecurity threats within a first domain. The first dataset includes labeled files associated with the first domain. The first trained machine learning model includes multiple layers, some of which are trainable. A second trained machine learning model is generated, via a transfer learning process, using (1) at least one trainable layer from the multiple trainable layers of the first trained machine learning model, and (2) a second dataset different from the first dataset. The second dataset includes labeled files associated with a second domain. The first domain has a different syntax, different semantics, and/or a different structure than that of the second domain. The second trained machine learning model (e.g., a deep neural network model) is then available for use in inferring cybersecurity-oriented properties of the file in the second domain and/or detecting cybersecurity threats in the second domain.

System and method for heterogeneous transferred learning for enhanced cybersecurity threat detection

Churn-aware training of a classifier which reduces the difference between predictions of two different models, such as a prior generation of a classification model and a subsequent generation. A second dataset of labelled data is scored on a prior generation of a classification model, wherein the prior generation was trained on a first dataset of labelled data. A subsequent generation of a classification model is trained with the second dataset of labelled data, wherein in training of the subsequent generation, weighting of at least some of the labelled data in the second dataset, such as labelled data threat yielded an incorrect classification, is adjusted based on the score of such labelled data in the prior generation.

Churn-aware machine learning for cybersecurity threat detection

A method for performing cyber-security analysis includes storing a semantic graph with nodes representing monitored computer-based entities, and edges representing monitored relationships. Each edge has an associated tally. A set of threat scores associated with multiple computer-based entities is stored in the memory. The semantic graph is updated in response to receiving event data. The updating includes decomposing the event data into a set of entities and a set of associated relationships, updating the tally of one of the edges based on the set of relationships, modifying an alert attribute of a monitored computer-based entity when the event data includes an applicable alert, and modifying a threat score of at least one computer-based entity based on the event data when the event data includes an applicable alert, to define a set of modified threat scores. The updated semantic graph is monitored for cyber-security risks within the multiple computer-based entities.

A system for detecting whether a file including content is associated with a cyber-attack is described. The content may include an executable file for example. The system includes an intelligence-driven analysis subsystem and a computation analysis subsystem. The intelligence-driven analysis subsystem is configured to (i) receive the file, (ii) inspect and compute features of the file for indicators associated with a cyber-attack, and (iii) produce a first output representing the detected indicators. The computational analysis subsystem includes an artificial neural network to (i) receive a network input being a first representation of at least one section of binary code from the file as input, and (ii) process the first representation of the section to produce a second output. The first output and the second output are used in determination a classification assigned to the file.

Growing community of inventors

Cary, NC, United States of America

Scott Eric Coull