Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, to process audio and visual data. One of the methods includes receiving, from a mobile device, image data and audio data, wherein the image data and audio data were obtained within the same time frame. The method includes processing the image data to identify a first set of facts. The method includes processing the audio data to identify a second set of facts. The method includes processing meta-data associated with the image data to identify a third set of facts. The method also includes performing at least one process using the first set of facts, the second set of facts, and the third set of facts.

Abstract: Systems and methods ingest extensive data regarding real-time transactions to determine correlation between a variety of parameters and possible fraud events. Possible fraud can be predicted, and notifications delivered to a fraud team. Correlations can be discovered using unsupervised machine learning, and particular systems and methods can utilize natural language processing to both receive and provide information concerning fraud risk and countermeasures.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, to process audio and visual data. One of the methods includes receiving, from a mobile device, image data and audio data, wherein the image data and audio data were obtained within the same time frame. The method includes processing the image data to identify a first set of facts. The method includes processing the audio data to identify a second set of facts. The method includes processing meta-data associated with the image data to identify a third set of facts. The method also includes performing at least one process using the first set of facts, the second set of facts, and the third set of facts.

Abstract: Systems and methods ingest extensive data regarding real-time transactions to determine correlation between a variety of parameters and possible fraud events. Possible fraud can be predicted, and notifications delivered to a fraud team. Correlations can be discovered using unsupervised machine learning, and particular systems and methods can utilize natural language processing to both receive and provide information concerning fraud risk and countermeasures.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, to process audio and visual data. One of the methods includes receiving, from a mobile device, image data and audio data, wherein the image data and audio data were obtained within the same time frame. The method includes processing the image data to identify a first set of facts. The method includes processing the audio data to identify a second set of facts. The method includes processing meta-data associated with the image data to identify a third set of facts. The method also includes performing at least one process using the first set of facts, the second set of facts, and the third set of facts.

Abstract: Techniques are described for determining and performing actions associated with a vehicle based on an analysis of sensor data that describes a state of component(s) of the vehicle. Sensor data is generated by sensor(s) on a vehicle and/or external to the vehicle. The sensor data is analyzed to determine a current state of vehicle component(s), such as whether a component is damaged or otherwise in need of repair or maintenance. The analysis of the sensor data may be performed by a rules engine that applies a set of rules indicating what action(s) are to be performed based on certain detected states of the vehicle component(s) indicated by the sensor data. The rule(s) may also indicate instances in which action(s) can be performed automatically or autonomously in response to detecting vehicle component state(s), or whether authorization is to be requested from an owner, operator, or other individual associated with the vehicle.

Abstract: Click-to-callback provides an icon or link on a webpage that a user can select or click on. When the icon or link is clicked or otherwise selected, the user is provided with a list of his phone numbers to choose from, or a blank field to enter an alternate number, for an immediate return call from a representative of the entity (e.g., the company) associated with the webpage. When this request is submitted, the representative desirably calls the user back immediately (e.g., within about 10 seconds) to provide assistance. Thus, the user remains on the website, and gets the help he desires to continue on the website, which may, for example, lead to the user making an online purchase.

Abstract: A computer-implemented method for monitoring a condition of a person includes receiving, at a computerized device, at least one signal from a condition sensor and determining if a condition is an emergency condition of a user based on the at least one signal.