Patent application number | Description | Published |
20080319932 | CLASSIFICATION USING A CASCADE APPROACH - A system and method that facilitates and effectuates optimizing a classifier for greater performance in a specific region of classification that is of interest, such as a low false positive rate or a low false negative rate. A two-stage classification model can be trained and employed, where the first stage classification is optimized over the entire classification region and the second stage classifier is optimized for the specific region of interest. During training the entire set of training data is employed by a first stage classifier. Only data that is classified by the first stage classifier or by cross validation to fall within a region of interest is used to train the second stage classifier. During classification, data that is classified within the region of interest by the first classification is given the first stage classifier's classification value, otherwise the classification value for the instance of data from the second stage classifier is used. | 12-25-2008 |
20090006575 | Detection and Removal of Undesirable Items in a Data Processing Environment - Functionality is described for addressing a threat to the security of a user device that utilizes a network-accessible service. The functionality operates by assessing the likelihood that the user device is infected by the undesirable item. When the user device makes a request to access the network-accessible service, the functionality can interact with the user device in a manner that is governed by the assessed likelihood that the user device is infected by the undesirable item. | 01-01-2009 |
20090265786 | AUTOMATIC BOTNET SPAM SIGNATURE GENERATION - A framework may be used for generating URL signatures to identify botnet spam and membership. The framework may take a set of unlabeled emails as input that are grouped based on URLs contained within the emails. The framework may return a set of spam URL signatures and a list of corresponding botnet host IP addresses by analyzing the URLs within the emails that are contained within the groups. Each URL signature may be in the form of either a complete URL string or a URL regular expression. The signatures may be used to identify spam emails launched from botnets, while the knowledge of botnet host identities can help filter other spam emails also sent by them. | 10-22-2009 |
20100162391 | Online Risk Mitigation - Online risk mitigation techniques are described. In an implementation, a service is queried for a reputation associated with an object from an online source in response to selection of the object. A backup of a client that is to receive the object is stored prior to obtaining the object when the reputation does not meet a threshold reputation level. | 06-24-2010 |
20110191832 | RESCUING TRUSTED NODES FROM FILTERING OF UNTRUSTED NETWORK ENTITIES - Network entities controlling a set of nodes may vary by trustworthiness, such as tolerance for nodes that send spam, distribute malware, or perform denial-of-service attacks. A device receiving such activities may identify a trust rating of the network entity and apply appropriately stringent filtering (such as spam evaluation) to activities received from nodes controlled by the network entity. However, a poor trust rating of a network entity may subject a legitimate node controlled by the network entity to inefficiently or unfairly stringent activity filtering. Instead, the device may evaluate the activities of a particular node, assign a trust rating to the node, and if the trust rating of the node is higher than the trust rating of the network entity, apply less stringent activity filtering to the activities of the node, thereby “rescuing” the node from the more stringent activity filtering applied to the other nodes of the network entity. | 08-04-2011 |
20110191847 | ACTIVITY FILTERING BASED ON TRUST RATINGS OF NETWORK ENTITIES - The filtering of activities generated by nodes of a network while interacting with a device may be performed by evaluating the desirability of the activities (e.g., a spam or not-spam determination of email messages sent by the node) and assigning a trust rating to the node. However, nodes are often identified by network address, and an operator of a node sending undesirable activities may reassign the network address of the node in order to avoid heavy filtering. Instead, nodes may be identified as being controlled by a network entity (e.g., an autonomous system identified in a border gateway protocol routing table.) The network entity is assigned a network entity trust rating based on the trust ratings of the nodes controlled thereby, and an appropriate level of activity filtering based on the network entity trust rating may be selected for subsequent activities received from all nodes controlled by the network entity. | 08-04-2011 |
20110247061 | COMPUTATION TO GAIN ACCESS TO SERVICE - Access to some aspect of a service may be limited until a user has invested in performing some amount of computation. Legitimate users typically have excess cycles on their machines, which can be used to perform computation at little or no cost to the user. By contrast, computation is expensive for for-profit internet abusers (e.g., spammers). These abusers typically use all of their computing resources to run “bots” that carry out their schemes, so computation increases the abuser's cost by forcing him or her to acquire new computing resources or to rent computer time. Thus, the providers of free services (e.g., web mail services, blogging sites, etc.), can allow newly registered users to use some limited form of the service upon registration. However, in order to make more extensive use of the service, the user can be asked to prove his legitimacy by investing in some amount of computation. | 10-06-2011 |
20110314537 | AUTOMATIC CONSTRUCTION OF HUMAN INTERACTION PROOF ENGINES - Human Interaction Proofs (“HIPs”, sometimes referred to as “captchas”), may be generated automatically. An captcha specification language may be defined, which allows a captcha scheme to be defined in terms of how symbols are to be chosen and drawn, and how those symbols are obscured. The language may provide mechanisms to specify the various ways in which to obscure symbols. New captcha schemes may be generated from existing specifications, by using genetic algorithms that combine features from existing captcha schemes that have been successful. Moreover, the likelihood that a captcha scheme has been broken by attackers may be estimated by collecting data on the time that it takes existing captcha schemes to be broken, and using regression to estimate the time to breakage as a function of either the captcha's features or its measured quality. | 12-22-2011 |
20120259929 | GEO-DATA SPAM FILTER - Geo-data spam filters are described. In one or more implementations, origin data and language data of a message are evaluated to establish a score for the message indicating a likelihood that the message is spam. The evaluation includes comparing the origin data and the language data to ranked lists indicating message origins and languages with which a respective message recipient interacts positively and ranked lists indicating message origins and languages with which the respective recipient interacts negatively. Interactions of the respective recipient with previously sent messages may be tracked to form these lists. Based on the score established by evaluating the origin data and the language data of the message, the message is filtered for delivery. | 10-11-2012 |
20130145384 | USER INTERFACE PRESENTING AN ANIMATED AVATAR PERFORMING A MEDIA REACTION - This document describes techniques and apparatuses enabling a user interface for presenting a media reaction. The techniques receive media reactions of a person to a media program, such as the person laughing at one point of a comedy show, then smiling at another point, and then departing at a third point. The techniques may present these and other media reactions in a user interface through which a user may interact. | 06-06-2013 |
20130145385 | CONTEXT-BASED RATINGS AND RECOMMENDATIONS FOR MEDIA - This document describes techniques and apparatuses enabling context-based ratings and recommendations for media. The techniques and apparatuses can build and continually improve the predictive accuracy of a user's reaction history based on a context in which the user's reactions to media are sensed. Further, the techniques and apparatuses may take into account a current context of a user when a request for a rating or recommendation is made. Based on the user's reaction history and the user's current context, the techniques and apparatuses may provide accurate ratings and recommendations. | 06-06-2013 |
20130152113 | DETERMINING AUDIENCE STATE OR INTEREST USING PASSIVE SENSOR DATA - This document describes techniques and apparatuses for determining an audience's state or interest using passive sensor data. The techniques receive sensor data that measures an audience during presentation of a media program. The techniques, based on this sensor data, then determine a state or interest of the audience members during the media program. | 06-13-2013 |
20130268954 | CONTROLLING A MEDIA PROGRAM BASED ON A MEDIA REACTION - This document describes techniques and apparatuses for controlling a media program based on a media reaction. In some embodiments, the techniques pause, mute, or stop a media program when a user leaves the room, when a user in the room is talking or is otherwise not paying attention to the program, or when a child walks into the room. | 10-10-2013 |
20130268955 | HIGHLIGHTING OR AUGMENTING A MEDIA PROGRAM - This document describes techniques and apparatuses for highlighting or augmenting a media program. The techniques and apparatuses can build a media program highlighting another media program based on media reactions to portions of that other media program. The techniques and apparatuses may also or instead augment a media program based on media reactions to portions of that media program. | 10-10-2013 |
20130298146 | DETERMINING A FUTURE PORTION OF A CURRENTLY PRESENTED MEDIA PROGRAM - This document describes techniques and apparatuses for determining a future portion of a currently presented media program. The techniques and apparatuses can receive current media reactions of one or many people to a currently presented media program and determine later portions to present in the media program based on the media reactions. In some embodiments, for example, a program can be presented live, reactions can be received during the live presentation, and the program altered on-the-fly and in real time based on those reactions. Further, the alterations can be general or tailored to a group or a particular person. | 11-07-2013 |
20130298158 | ADVERTISEMENT PRESENTATION BASED ON A CURRENT MEDIA REACTION - This document describes techniques and apparatuses enabling advertisement presentation based on a current media reaction. The techniques and apparatuses can receive a current media reaction of a user watching a media program and, based on this current media reaction, determine which advertisement is likely to be effective. Further, the techniques and apparatuses may inform advertisers of a current media reaction thereby enabling the advertisers to bid on a right to present an advertisement based on that reaction. By so doing, costs for advertisements may more-accurately reflect the value of the time in which they are presented and advertisements may be more effective. | 11-07-2013 |
20140259104 | AUTOMATIC CONSTRUCTION OF HUMAN INTERACTION PROOF ENGINES - Human Interaction Proofs (“HIPs”, sometimes referred to as “captchas”), may be generated automatically. An captcha specification language may be defined, which allows a captcha scheme to be defined in terms of how symbols are to be chosen and drawn, and how those symbols are obscured. The language may provide mechanisms to specify the various ways in which to obscure symbols. New captcha schemes may be generated from existing specifications, by using genetic algorithms that combine features from existing captcha schemes that have been successful. Moreover, the likelihood that a captcha scheme has been broken by attackers may be estimated by collecting data on the time that it takes existing captcha schemes to be broken, and using regression to estimate the time to breakage as a function of either the captcha's features or its measured quality. | 09-11-2014 |