Abstract: Distinguishing between functional tracking domains and nonfunctional tracking domains on a host web page. In particular, a list of known tracking domains that load content into host web pages may be received. This list of tracking domains may include tracking domains that are functional and tracking domains that are nonfunctional. The tracking domains that are functional may be determined by evaluating various behaviors and characteristics of the tracking domains. Once functional tracking domains have been determined, these functional tracking domains may be allowed, and other tracking domains may be blocked from loading content onto host web pages thereby preserving the functionality of the web pages.

Abstract: The disclosed computer-implemented method for utilizing user profile data to protect against phishing attacks may include (i) detecting a target user profile associated services accessed by a network-based application, (ii) determining identifiers associated with each of the services, (iii) extracting, for each of the identifiers, feature vectors describing exploitable screen elements in the network-based application associated with phishing attacks, (iv) updating, based on the feature vectors, previously extracted feature vectors in a data repository storing additional profiles for other users associated with the services, (v) predicting, utilizing a machine-learning model, phishing attack threats for target profile user based on a similarity with the additional profiles, and (vi) performing a security action that protects against the phishing attack threats. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: Systems and methods for identifying and removing a tracking capability from an external domain that performs a tracking activity on a host web page. Tracking capabilities of an external domain may be removed by altering web requests and/or responses to API calls. Once these tracking capabilities of the external domain have been removed, the altered web requests and/or altered responses to API calls may be transmitted to a web browser and/or entity making the API call thereby protecting user privacy while allowing the external domain to interact with the host web page.

Abstract: Identifying and removing a tracking capability from an external domain that performs a tracking activity on a host web page. Tracking capabilities of an external domain may be removed by altering web requests and/or responses to API calls. Once these tracking capabilities of the external domain have been removed, the altered web requests and/or altered responses to API calls may be transmitted to a web browser and/or entity making the API call thereby protecting user privacy while allowing the external domain to interact with the host web page.

Abstract: The disclosed computer-implemented method for identifying privacy leakage information may include (1) identifying, at the computing device, at least one informative word in a digital text and (2) performing a security action that identifies privacy leakage information, where the security action includes (A) determining, for at least one identified informative word, a type of privacy leakage and a respective confidence score indicating a probability the identified informative word causes the type of privacy leakage, (B) determining, using the respective confidence score, a combined confidence score for each respective element within a level of detail to display, and (C) displaying, on a display device, the combined confidence score for each respective element within the level of detail to display. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: Methods and systems are provided for monitoring private information exposure. One example method generally includes detecting, at a computing device, a potentially exposing post on a publicly-accessible server and determining, by an exposure model operated by the computing device, an exposure chance for the potentially exposing post. The method further includes transmitting an initial notification of the potentially exposing post from the computing device to a client device based on the exposure chance from the exposure model and detecting an update event associated with the potentially exposing post. The method further includes updating the exposure model based on the update event and transmitting a personalized notification from the computing device to the client device based on the update event.

Abstract: The disclosed computer-implemented method for generating user profiles may include (i) analyzing a data set of user profiles for services, (ii) detecting a measurement of obfuscation that was applied to a specific attribute across multiple user profiles for a specific service, (iii) applying the measurement of obfuscation to true data for a new user by fuzzing the true data to create a fuzzed value, and (iv) generating automatically a new user profile for the specific service by populating the specific attribute within the new user profile with the fuzzed value. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: Determining security vulnerabilities of smart home devices.

Abstract: The disclosed computer-implemented method for fingerprinting devices may include (i) detecting that a new device has attempted to connect to a network gateway, (ii) attempting to fingerprint the new device as an instance of a known candidate device type by (a) transmitting to the new device, from a security application, a set of network messages that mimic network messages that a second application is configured to transmit to instances of the known candidate device type and (b) confirming, by the security application based on a response from the new device to the set of network messages, that the new device is the instance of the known candidate device type, and (iii) performing a security action to protect a network corresponding to the network gateway based on confirming that the new device is the instance of the known candidate device type. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: A method for anomaly detection using grouping-based graph summarization is performed by one or more computing devices, each including one or more processors. The method includes identifying a first graph from a stream of incoming graphs, generating a first summary of the first graph, classifying the first summary in a first category, identifying a second graph from the stream of incoming graphs, generating a second summary of the second graph, comparing the first summary with the second summary, mapping the first summary and the second summary to the first category upon determining that the comparing indicates a similarity between the first summary and the second summary satisfies a graph similarity threshold, analyzing a frequency of graphs being mapped to graph categories, the graph categories including at least the first category, and detecting an anomaly in one of the graph categories based at least in part on the analysis.

Abstract: The disclosed computer-implemented method for controlling uploading of potentially sensitive information to the Internet may include (i) loading, at the computing device, at least a portion of a webpage and (ii) performing a security action including (A) converting, at the computing device, components of the webpage from an online status to an offline status, (B) receiving a sensitive information input to a respective offline component of the webpage, (C) converting, based on a stored user preference and in response to receiving the sensitive information input, the respective offline component to the online status, (D) buffering an outgoing network request comprising the sensitive information input, (E) receiving an approval input indicating approval to transmit the potentially sensitive information to the Internet, and (F) releasing the outgoing network request in response to receiving the approval input. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: Methods and systems are provided for detecting privacy leakage risks in text. One example method generally includes receiving, at a computing device from a client device, a keyword and generating, by the computing device, a combined query comprising the keyword and a sensitive topic query associated with a sensitive topic. The method further includes transmitting the keyword from the computing device to a search engine and transmitting the combined query from the computing device to the search engine. The method further includes receiving, at the computing device from the search engine, a number of search results for the keyword and receiving, at the computing device from the search engine, a number of search results for the combined query. The method further includes determining, by the computing device, a confidence score and transmitting the confidence score from the computing device to the client device.

Abstract: Utility driven graph summarization for use in detecting and preventing malicious computer application. In one embodiment, a method may include receiving a graph comprising a plurality of nodes and a plurality of edges, prioritizing each of the plurality of nodes by way of assigning a relative importance value to each node of the plurality of nodes, combining at least two nodes of the plurality of nodes into a supernode based at least on the relative importance value of each node, calculating a utility penalty value for creating a superedge between the supernode and a node neighboring the supernode, creating the superedge between the supernode and the node neighboring the supernode if the utility penalty value satisfies a pre-determined penalty threshold, calculating a utility level based at least in part on creating the supernode and the superedge, and repeating the method until the calculated utility level satisfies a pre-determined threshold.

Abstract: Securing a network device from a malicious embedded script hosted on a third-party domain. In one embodiment, a method may include detecting an attempt by a browser executing on a network device to load a webpage that embeds a reference to a script hosted on a third-party domain, compiling a list of domains that host webpages that embed references to the script hosted on the third-party domain, identifying reputation scores for the domains in the list of domains, generating a risk score for the script based on the identified reputation scores, determining that the script is malicious based on the generated risk score being above a threshold risk score, and, in response to determining that the script is malicious, performing a security action on the network device that secures the network device from the malicious script.

Abstract: User-directed identity verification over a network. In one embodiment, a method may include receiving identity information of a user, storing the identity information, calculating a confidence score for the identity information that indicates a level of confidence that the identity information belongs to the user, receiving a request for the identity information of the user, or for confirmation of possession of the identity information of the user, to be sent to an organization, sending an authorization request for the user to authorize the organization to access the identity information or to receive the confirmation of possession of the identity information, receiving a grant of authorization for the organization to access the identity information or to receive the confirmation of possession of the identity information, sending the identity information, or the confirmation of possession of the identity information, and the confidence score for the identity information.

Abstract: Building a ground truth dataset for a machine learning-based security application. In one embodiment, a method may include identifying a set of network devices to add to a ground truth dataset. The method may also include, for each network device in the set of network devices, identifying a potentially malicious application stored on the network device, analyzing behavior of the potentially malicious application to determine whether the potentially malicious application has behaved maliciously, and if so, adding the network device to the ground truth dataset as an infected device or, if not, adding the network device to the ground truth dataset as a clean device. The method may further include training a machine learning classifier of a security application using the ground truth dataset, making a security action decision using the machine learning classifier, and performing a security action on a computer system based on the security action decision.

Abstract: The disclosed computer-implemented method for identifying privacy leakage information may include (1) identifying, at the computing device, at least one informative word in a digital text and (2) performing a security action that identifies privacy leakage information, where the security action includes (A) determining, for at least one identified informative word, a type of privacy leakage and a respective confidence score indicating a probability the identified informative word causes the type of privacy leakage, (B) determining, using the respective confidence score, a combined confidence score for each respective element within a level of detail to display, and (C) displaying, on a display device, the combined confidence score for each respective element within the level of detail to display. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: Securing network devices by forecasting future security incidents for a network based on past security incidents. In one embodiment, a method may include constructing past inside-in security features for a network, constructing past outside-in security features for the network, and employing dynamic time warping to generate a similarity score for each security feature pair in the past inside-in security features, in the past outside-in security features, and between the past inside-in security features and the past outside-in security features. The method may further include generating a Coupled Gaussian Latent Variable (CGLV) model based on the similarity scores, forecasting future inside-in security features for the network using the CGLV model, and performing a security action on one or more network devices of the network based on the forecasted future inside-in security features for the network.

Abstract: Preventing a malicious computer application from executing in a computing environment. In one embodiment, a method may include identifying a base graph, identifying a perturbed graph, determining an importance of each of the edges in the base graph using an Edge Current-Flow Based Betweenness Centrality (ECFBBC) metric, identifying the edges in the base graph that match the edges in the perturbed graph, determining a utility value for the perturbed graph, determining whether the utility value is above a threshold utility value, in response to determining that the utility value is above the threshold utility value, employing the perturbed graph to analyze a computer application and determine that the computer application is malicious, and performing a security action on the malicious computer application to prevent the malicious computer application from executing in a computing environment.