Abstract: Employing a relatively simple machine learning classifier to explain evidence that led to a security action decision by a relatively complex machine learning classifier. In one embodiment, a method may include identifying training data, training a relatively complex machine learning classifier (MLC) using the training data, making a first security action decision using the relatively complex MLC, performing a security action on a network device based on the first security action decision, training a relatively simple MLC using the training data or using a subset of the training data, making a second security action decision using the relatively simple MLC, extracting evidence that led to the second security action decision by the relatively simple MLC, and explaining the first security action decision by the relatively complex MLC using the evidence extracted from the relatively simple MLC.

Abstract: The disclosed computer-implemented method for generating training documents used by classification algorithms may include (i) identifying a set of training documents used by a classification system to classify documents written in a first language, (ii) generating a list of tokens from within the training documents that indicate critical terms representative of classes defined by the classification system, (iii) translating the list of tokens from the first language to a second language, (iv) creating, based on the translated tokens, a set of simulated training documents that enables the classification system to classify documents written in the second language, and (v) classifying an additional document written in the second language based on the set of simulated training documents. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: Blocking malicious Internet content at an appropriate hierarchical level. In one embodiment, a method may include identifying evidence of security risks in hierarchical levels of an Internet hierarchy. The method may also include generating security risk scores for the hierarchical levels of the Internet hierarchy based on the evidence of security risks. The method may further include identifying a security risk threshold. The method may also include identifying, as an appropriate blocking level, the highest hierarchical level of the Internet hierarchy having a security risk score at or above the security risk threshold. The method may further include blocking a network device from accessing Internet content in the Internet hierarchy at or below the appropriate blocking level.

Abstract: Blocking malicious Internet content at an appropriate hierarchical level. In one embodiment, a method may include identifying evidence of security risks in hierarchical levels of an Internet hierarchy. The method may also include generating security risk scores for the hierarchical levels of the Internet hierarchy based on the evidence of security risks. The method may further include identifying a security risk threshold. The method may also include identifying, as an appropriate blocking level, the highest hierarchical level of the Internet hierarchy having a security risk score at or above the security risk threshold. The method may further include blocking a network device from accessing Internet content in the Internet hierarchy at or below the appropriate blocking level.

Abstract: According to one embodiment, a method for predicting the trustworthiness of a particular website comprising receiving information about a plurality of websites and constructing a hierarchy of groups from the received information, the hierarchy of groups comprising one or more tiers and each tier comprising one or more groups. The method further comprising receiving information about a particular website and predicting the trustworthiness of the particular website based on the hierarchy.

Abstract: According to one embodiment, a method for predicting the trustworthiness of a particular website comprising receiving information about a plurality of websites and constructing a hierarchy of groups from the received information, the hierarchy of groups comprising one or more tiers and each tier comprising one or more groups. The method further comprising receiving information about a particular website and predicting the trustworthiness of the particular website based on the hierarchy.

Abstract: Aspects of the disclosure can provide a face recognition method of increased robustness. The face recognition method can include receiving a target face image, initializing a registration and a block analysis of the target face image identifying one or more portions of the target face image for face recognition, informing a user of the registration and the block analysis, receiving one or more instructions from the user for operating on the registration and the block analysis, and identifying the target face image from a plurality of face images based on the registration and the block analysis of the target face image when the registration and the block analysis are approved by the user.

Abstract: A biometric speaker-identification apparatus is disclosed that generates ordered speaker-identity candidates for a probe based on prototypes. Probe match scores are clustered, and templates that correspond to clusters having top M probe match scores are compared with the prototypes to obtain template-prototype match scores. The probe is also compared with the prototypes, and those templates corresponding to template-prototype match scores that are nearest to probe-prototype match scores are selected as speaker-identity candidates. The speaker-identity candidates are ordered based on their similarity to the probe.

Abstract: A prototype biometric identification system is disclosed that indexes a biometric corpus into indexed-corpuses using a set of P prototypes before searching for a probe in a search corpus constructed based on the indexed-corpus. The system may index the biometric corpus based on the prototypes directly or based on prototype-typicality scores.

Abstract: A biometric speaker-identification apparatus is disclosed that generates ordered speaker-identity candidates for a probe based on prototypes. Probe match scores are clustered, and templates that correspond to clusters having top M probe match scores are compared with the prototypes to obtain template-prototype match scores. The probe is also compared with the prototypes, and those templates corresponding to template-prototype match scores that are nearest to probe-prototype match scores are selected as speaker-identity candidates. The speaker-identity candidates are ordered based on their similarity to the probe.

Abstract: A prototype biometric identification system is disclosed that indexes a biometric corpus into indexed-corpuses using a set of P prototypes before searching for a probe in a search corpus constructed based on the indexed-corpus. The system may index the biometric corpus based on the prototypes directly or based on prototype-typicality scores.

Abstract: Aspects of the disclosure can provide a face recognition method of increased robustness. The face recognition method can include receiving a target face image, initializing a registration and a block analysis of the target face image identifying one or more portions of the target face image for face recognition, informing a user of the registration and the block analysis, receiving one or more instructions from the user for operating on the registration and the block analysis, and identifying the target face image from a plurality of face images based on the registration and the block analysis of the target face image when the registration and the block analysis are approved by the user.

Abstract: The present invention includes a method and system for image compression using block size heuristics. A method for motion searching a video frame is disclosed including iteratively decreasing block size until a rate-distortion (RD) has been minimized. A method for compressing motion video images is disclosed. Additionally, a system for transmitting and receiving video images is disclosed. The system may be a video conferencing system.

Abstract: The present invention includes a method and system for improved diamond motion search. A method for diamond motion searching a video frame is disclosed which includes predicting the maximum distance that a macroblock may have moved. This maximum distance provides a maximum range in which to consider searching. This “predicted search range” may be used to make assumptions on whether to expect high motion. If high motion is anticipated, the diamond search may be seeded using a large circular pattern for determining a start location and to avoid becoming lost in local minima and then proceeding with the large diamond pattern for motion searching. A method for compressing motion video images is also disclosed. Additionally, a system for transmitting and receiving video images is disclosed. The system for transmitting and receiving video images may be a video conferencing system.

Abstract: The present invention includes a method and system for image compression using block size heuristics. A method for motion searching a video frame is disclosed including iteratively decreasing block size until a rate-distortion (RD) has been minimized. A method for compressing motion video images is disclosed. Additionally, a system for transmitting and receiving video images is disclosed. The system may be a video conferencing system.