Netskope, Inc.

Zscaler, Inc.

Dong Guo

Yi Zhang

Yihua Liao

Krishna Narayanaswamy

Siying Yang

Rongrong Tao

Howie Xu

Rex Shang

Changsha Ma

Dianhuan Lin

Xinjun Zhang

Douglas A Koch

Jason B Bryslawskyj

Jun Ou

Hongbo Yang

A network security system that classifies webpages and uses the classification of the webpages to enforce relevant security policies is disclosed. To classify a webpage, a screenshot of at least a portion of the webpage is captured. An embedding engine generates a subject image embedding of the screenshot. The subject image embedding is classified by an image classifier that includes an index of training image embeddings each having a label of their classification and an associated approximate nearest neighbors model trained to identify the label of the closest training image embedding to the subject image embedding and a score representing their similarity. The webpage is classified based at least in part on the score and the label from the image classifier. The network security system applies security policies to requests from client devices that identify the webpage as its destination based at least in part on the webpage classification.

Webpage categorization based on image classification of webpage screen capture

Disclosed is a method of building a customized deep learning (DL) stack classifier to detect organization sensitive data in images, referred to as image-borne organization sensitive documents, and protecting against loss of the image-borne organization sensitive documents, including distributing a trained feature map extractor stack, with stored parameters, configured to allow the organization to extract from image-borne organization sensitive documents, feature maps that are used to generate updated DL stacks and to save non invertible feature maps derived from the images, and ground truth labels for the image. Also included is receiving organization-specific examples including the non-invertible feature maps extracted from the organization-sensitive documents and the ground truth labels and using the received organization-specific examples to update a customer-specific DL stack classifier. Further included is sending the customer-specific DL stack classifier to the organization.

Detecting organization sensitive data in images via customized deep learning classifier

Disclosed is technology for choosing between alternative category labels tentatively assigned to tens of thousands of webpages by a classifier ensemble running on processors, applying the classifier ensemble with a sensitive category classifier, a non-sensitive category classifier, a title and metadata classifier and a heuristic classifier to tens of thousands of webpages. Also disclosed is applying a post processor to outputs of the classifier ensemble and tentatively assigning at least two category labels for non-sensitive categories for webpages; two category labels, automatically determining that at least one but not all of the tentatively assigned category labels is a general label and de-assigning the general label; saving the category label that is not de-selected to the webpage; and distributing the assigned category labels for at least some of the tens of thousands of webpages for use in controlling access to webpages by users on user systems protected using the assigned labels.

Machine learning-based URL categorization system with selection between more and less specific category labels

Systems and methods include obtaining file identifiers associated with files in production data; obtaining lab data from one or more public repositories of malware samples based on the file identifiers for the production data; and utilizing the lab data for training a machine learning process for classifying malware in the production data. The obtaining file identifiers can be based on monitoring of users associated with the files, and only the file identifiers are maintained based on the monitoring. The lab data can include samples from the one or more public repositories matching the corresponding file identifiers for the production data. The lab data can include samples from the one or more public repositories that have features closely related to features of the production data.

Building a Machine Learning model without compromising data privacy

Disclosed is a method of building a customized deep learning (DL) stack classifier to detect organization sensitive data in images, referred to as image-borne organization sensitive documents, and protecting against loss of the image-borne organization sensitive documents, including distributing a trained feature map extractor stack with stored parameters to an organization, under the organization's control, configured to allow the organization to extract from image-borne organization sensitive documents, feature maps that are used to generate updated DL stacks, without the organization forwarding images of organization-sensitive training examples, and to save non invertible feature maps derived from the images, and ground truth labels for the image. Also included is receiving organization-specific examples including the non-invertible feature maps extracted from the organization-sensitive documents and the ground truth labels and using the received organization-specific examples to generate a customer-specific DL stack classifier. Further included is sending the customer-specific DL stack classifier to the organization.

Customized deep learning classifier for detecting organization sensitive data in images on premises

Growing community of inventors

San Jose, CA, United States of America

Dong Guo