Abstract: A system for obtaining user parameters of e-commerce users to auto complete checkout forms includes one or more processors configured to monitor user interactions of a user with a merchant website on a user device; determine an intent to purchase one or more items based of the monitoring; determine an unfilled template of a checkout form associated with the merchant website based on previously stored website information, wherein the unfilled template of the checkout form comprises a plurality of checkout form fields; determine one or more user parameters associated with the user device; apply a machine learning algorithm to predict a geolocation of the user device based on the one or more user parameters; and auto fill the at least one of the plurality of checkout form fields in the unfilled template of the checkout form based on the predicted geolocation.

Abstract: In some implementations, a user device may access a media presentation that includes metadata describing related content item(s). The user device viewing the media presentation is allowed to access content related to portions of the media presentation at times appropriate for the particular related content item(s). The related content item(s) may be provided automatically or based on user input triggering download or copying of a particular related content item, such as to a clipboard of the user device. A computing device may generate a media presentation that includes related content item(s) as metadata in some implementations. The media presentation may be generated by recording a live presentation, assembling one or more media portions, and/or obtaining a complete media presentation and modifying the media presentation to add the related content item(s) and when, while presenting the media presentation on a user device, to allow access to the related content item(s).

Abstract: A computing system accesses crime incident data from two or more organization systems. The crime incident data includes a first set of inputs associated with a plurality of crime incident groupings and a second set of inputs associated with incident data for incidents associated with each of the two or more organization systems. The computing system preprocesses the crime incident data to remove incidents that include suspect identifiers not present in at least two or more organization system. The computing system analyzes the preprocessed crime incident data to identify links between incidents across two or more organization systems using a trained crime linking model. The computing system generates a link between a first incident at a first organization system of the two or more organization systems and a second incident at a second organization system based on the analyzing.

Abstract: Examples of techniques for constructing, evaluating, and improving a search string for retrieving images are disclosed. In one example implementation according to aspects of the present disclosure, a computer-implemented method includes receiving, by a processing device, a plurality of images as search results returned based at least in part on a search string for an item in the form of a tuple including an item class, an action and an actor. The method further includes determining, by the processing device, whether the search string is effective at indicating a common item use based on image similarity. The method further includes, based at least in part on determining that the search string is ineffective at indicating the item use, generating, by the processing device, an alternative search string.

Abstract: Methods and devices related to a single message management platform are described. In an example, a method can include receiving a first message at a first processing resource via a first application, receiving a second message at the first processing resource via a second application, receiving the first message and the second message at a memory, storing the first message and the second message in the memory, receiving a command to open a single message management platform, receiving the first message and the second message at the single message management platform, and organizing the first message and the second message at the single message management platform.

Abstract: A system and methods for zero-step customer proximity detection using mobile device low emissions beacons, which uses a customer mobile device, a single or plurality of BLUETOOTH™ low-emissions (BTLE) beacons, a network, a server, and an application that may operate separately on a mobile device and the single or plurality of BLUETOOTH™ low-emissions (BTLE) beacons, which allows businesses to detect known customers that come within a specified proximity of any connected BTLE beacons, and provide interactions with them on a zero-step and socially distanced basis, track their proximity to other customers and staff using similar applications, manage table assignments, offer deals or provide incentives to enter or patronize a business establishment.

Abstract: Implementations provide for use of spherical random features for polynomial kernels and large-scale learning. An example method includes receiving a polynomial kernel, approximating the polynomial kernel by generating a nonlinear randomized feature map, and storing the nonlinear feature map. Generating the nonlinear randomized feature map includes determining optimal coefficient values and standard deviation values for the polynomial kernel, determining an optimal probability distribution of vector values for the polynomial kernel based on a sum of Gaussian kernels that use the optimal coefficient values, selecting a sample of the vectors, and determining the nonlinear randomized feature map using the sampled vectors.

Abstract: The present disclosure relates to a tag-based font recognition system that utilizes a multi-learning framework to develop and improve tag-based font recognition using deep learning neural networks. In particular, the tag-based font recognition system jointly trains a font tag recognition neural network with an implicit font classification attention model to generate font tag probability vectors that are enhanced by implicit font classification information. Indeed, the font recognition system weights the hidden layers of the font tag recognition neural network with implicit font information to improve the accuracy and predictability of the font tag recognition neural network, which results in improved retrieval of fonts in response to a font tag query. Accordingly, using the enhanced tag probability vectors, the tag-based font recognition system can accurately identify and recommend one or more fonts in response to a font tag query.

Abstract: Technologies are described for facilitating the processing of requests for database operations within a database environment that includes one or more processing units and a memory. A utilization level of one or more computing resources of a database system is periodically determined. When a request for a database operation is received from a database client, it is determined, such as based on the utilization level, whether the database system has sufficient resources to process the request. If the database system does not have sufficient resources to process the request, the request is enqueued in a request queue for deferred execution. If it is later determined that the database system has sufficient resources to execute the request, the request can be dequeued, executed, and execution results returned to the database client.

Abstract: An apparatus, method, and a computer readable medium for attenuating visual artifacts in processed images. An annotated dataset of images to be processed by an image processing system is created. An adversarial control network is trained to operate as an image quality expert in classifying images. After the adversarial control network has been trained, the adversarial control network is used to supervise the image processing system on-the-fly.

Abstract: A method, system and computer program product for event summarization facilitated by emotions/reactions of people near an event location is disclosed. The method includes generating a query based at least in part on reaction information and at least in part on primary video metadata. Based on the query, at least one possible event summarization match for the one or more events is retrieved from a database.

Abstract: Provided is a shooting method for a shooting device, including steps of: starting the shooting device to begin shooting a video when the shooting device detects a face, and recording face feature information corresponding to the video; uploading the face feature information and the corresponding video to a server and saving the face feature information and the corresponding video according to an upload and save condition after the shooting is completed, wherein binding relation information between the face feature information and the corresponding video is generated; the server receiving retrieve information, retrieving and matching corresponding face feature information according to the retrieve information, and generating a matching result; and the server outputting, according to the matching result and the binding relation information, video information bound with the matched face feature information.

Abstract: Systems and methods are described for processing ingested data using an online machine learning algorithm as the data is being ingested. For example, the online machine learning algorithm can be an adaptive thresholding algorithm used to identify outliers in a moving window of data. As another example, the online machine learning algorithm can be a sequential outlier detector that detects anomalous sequences of logs or events. As another example, the online machine learning algorithm can be a sentiment analyzer that determines whether text has a positive, negative, or neutral sentiment. As another example, the online machine learning algorithm can be a drift detector that detects whether ingested data marks the start of a change in the distribution of a time-series.

Abstract: Embodiments of the present disclosure provide a method and apparatus for processing positioning data, a device, a storage medium and a vehicle. In the method according to the embodiments of the present disclosure, the vehicle is controlled to collect positioning data continuously in the five stages including a first stage for performing dynamic alignment, a second stage for traveling along a path of a character “8”, a third stage for collecting map data, a fourth stage for traveling along a path of the character “8”, and a fifth stage for performing the dynamic alignment, and processing of error convergence is performed on the positioning data collected in the third stage, according to the positioning data collected in the first stage, the second stage, the fourth stage and the fifth stage, to obtain the final positioning data after the error convergence.

Abstract: An apparatus comprising an antenna array comprising a plurality of antennas to receive a plurality of radar signals reflected by a plurality of objects responsive to a transmitted radar signal; a doppler measurement module to determine, for a first reflected radar signal of the plurality of reflected radar signals, a first doppler measurement indicating a velocity component based on a comparison of the first reflected radar signal to the transmitted radar signal; a phase offset measurement module to determine a first phase offset of the first reflected radar signal received at a first antenna of the plurality of antennas relative to a phase of the first reflected radar signal received at a reference antenna of the plurality of antennas; and a phase offset calibration module to determine, for the first antenna, a first phase offset calibration error based on the first doppler measurement and the first phase offset.

Abstract: A method and system are provided including at least one tabular data set, wherein each data set includes one or more attribute categories defining a record, each attribute category including one or more data values; an anonymization module; and an anonymization processor in communication with the anonymization module and operative to execute processor-executable process steps to cause the system to: receive a first data set at the anonymization module; determine at least one of the attribute categories in the first data set is a first-class identifier; generate a pseudonymous data set by removing the at least one first-class identifier attribute category from the first data set; determine at least two of the attribute categories in the pseudonymous data set are a second-class identifier; manipulate the at least two second-class attribute categories; and generate an anonymized data set based on the at least two manipulated second-class identifier attribute categories. Numerous other aspects are provided.

Abstract: In accordance with embodiments, there are provided mechanisms and methods for appending data to large data volumes in a multi-tenant store. These mechanisms and methods for appending data to large data volumes can enable embodiments to provide more reliable and faster maintenance of changing data. In an embodiment and by way of example, a method for appending data to large data volumes is provided. The method embodiment includes receiving new data for a database. The new data is written to a temporary log. The size of the log is compared to a threshold. Then the log is written to a data store, if the size of the log is greater than the threshold.

Abstract: In some examples, visible elements-based application testing may include accessing, based on an analysis of a display of a page associated with execution of an application, an identification of elements and locations of the identified elements within the page. According to an example, each element of the identified elements may include a character. A determination may be made as to whether a word from a test script for testing the application is present in a set that includes at least two of the identified elements by comparing distances between a specified location of the word on the page and the locations of the identified elements. Further, in response to a determination that the word is present in the set of the identified elements, an operation associated with the word may be executed to test the application.

Abstract: A system, method, and computer program to implement Public Key Infrastructure (“PKI”)-based access controls on a computing device independently of the operating system's user identification and authorization system. A software application on a standalone, non-networked computing device detects a “smart card” with a PKI certificate identifying the card holder. The identity of the card holder is verified against a registry stored on the computing device separately from the operating system's user identification and authorization system. Verification of the identity of the card holder results in setting the state of the computer system as “authorized” or “unauthorized.” In the unauthorized state, user input and viewing of information is prevented. The software application may thus identify individual smart card holders and provide secure access to data and resources, and track authorized and unauthorized access on computing devices that must remain in an “always logged in” state.

Abstract: Systems and method are described for receiving, by a computing platform, information about a first user of a shared mobility service, generating, based on the information about the first user, a first user profile comprising one or more first user attributes, comparing the one or more first user attributes to one or more second user attributes of a second user profile associated with a second user of the shared mobility service, and based on the comparison, causing a vehicle carrying the second user to pick up the first user.

Abstract: The present invention relates to communications methods and apparatus for determining whether a received call originated from an artificial intelligence system and the call handling procedures to be used on the received call based on whether or not the call was determined to have been originated from an artificial intelligence system. An exemplary method embodiment includes the steps of: receiving a first call; determining whether said first call originated from an artificial intelligence system; and when said first call is determined to have originated from an artificial intelligence system performing one or more artificial intelligence call handling operations.

Abstract: A radar detection method may include: transmitting a first radar signal in a field of view and receiving a second radar signal originated from reflections of the first radar signal in the field of view; generating a detection profile by processing the first and second radar signals, the detection profile representing intensities of the second radar signal as a function of positions in the field of view; and analyzing the detection profile to identify targets in the field of view. Analyzing the detection profile may include: using a first mode of analysis, with lower sensitivity, for first cycles, wherein the first mode of analysis is configured to detect a target entering the field of view; using a second mode of analysis, with higher sensitivity, for second cycles following the first cycles, wherein the second mode of analysis is configured to detect stay of the target in the field of view.

Abstract: Systems and methods are provided for obtaining sensor data comprising one or more entities and one or more attributes of the respective one or more entities. The systems and methods may be configured to infer one or more relationships between the respective one or more entities based on the one or more attributes, and update the sensor data based on the inferred one or more relationships.

Abstract: Embodiments of the specification provide an image search method, an apparatus, and a device. The method includes: obtaining an input image associated with an image search, wherein the input image includes a plurality of first text blocks; selecting a to-be-processed image from a target database, wherein the to-be-processed image includes a plurality of second text blocks; and generating a first graph structural feature based on the plurality of first text blocks; generating a second graph structural feature based on the plurality of second text blocks; determining that the first graph structural feature and the second graph structural feature satisfy a condition; and in response to determining that the first graph structural feature and the second graph structural feature satisfy the condition, outputting the to-be-processed image as a search result.

Abstract: Methods, systems, and computer-readable media for query language interoperability in a graph database are disclosed. Data elements are inserted into a graph database using one or more of a plurality of graph database query languages. The graph database query languages comprise a first graph database query language associated with a first data model and a second graph database query language associated with a second data model. The data elements are stored in the graph database using an internal data model that differs from the first and second data models. One or more of the data elements are retrieved from the graph database based at least in part on a query. The query is expressed using a different graph database query language than the graph database query language used to insert the one or more retrieved data elements.

Abstract: The present disclosure relates to systems and methods for analyzing and extracting data related to a structured proceeding, and for identifying, based on the analysis, at least one outcome associated with the structured proceeding. Embodiments provide for receiving data associated with a structured proceeding involving at least one party, the data including at least one docket entry, and analyzing, by an outcome location detector, the data to identify one or more docket entries in the at least one docket entry that are likely to include evidence of an outcome. Embodiments further include analyzing, by an outcome detector, the one or more docket entries determined to be likely to include evidence of an outcome to determine outcomes. The outcomes include at least one of a final outcome and at least one party outcome. The final outcome is associated with the structured proceeding overall, and the at least one party outcome is associated with a party of the at least one party that may have been terminated early.

Abstract: Receive a transaction generated by a user of a non-internet application; identify transaction life cycle steps of previous similar transactions; and generate a transaction risk score for the transaction using machine learning models and a blockchain record of the previous similar transactions. In response to the transaction risk score exceeding a threshold value, authenticate the transaction and the user using two-step authentication. The two-step authentication uses challenge/answer templates derived from the blockchain record of previous transactions.

Abstract: According to various embodiments of the present disclosure, an electronic device includes a display, a communication circuit, a processor operatively connected to the display and the communication circuit, and a memory operatively connected to the processor, wherein the memory is configured to store a web browser including a uniform resource locator (URL) address bar and at least one application program, wherein the memory is further configured to store instructions that, when executed by the processor, cause the electronic device to be configured to: display a first webpage including a search box, using the web browser, on the display, receive at least one search text through the search box and/or the URL address bar, transmit the at least one search text to a first external server associated with the first webpage via the communication circuit, receive a first search result associated with the search text from the first external server; display the first search result on the display, receive a user input fo

Abstract: Techniques and systems are described for analytics system entity resolution. Typed higher-order node combinations are determined within a dataset, and an amount of similarity between two arbitrary nodes within the dataset is determined based on the typed higher-order node combinations. The amount of similarity enables the digital analytics to accurately perform source resolution of portions of the dataset to a respective source, and may be utilized to control output of digital content to a client device.

Abstract: The invention provides for the evaluation of semantic closeness of a source data file relative to candidate data files. The system includes an artificial neural network and processing intelligence that derives a property vector from extractable measurable properties of a data file. The property vector is mapped to related semantic properties for that same data file and such that, during ANN training, pairwise similarity/dissimilarity in property is mapped, during towards corresponding pairwise semantic similarity/dissimilarity in semantic space to preserve semantic relationships. Based on comparisons between generated property vectors in continuous multi-dimensional property space, the system and method assess, rank, and then recommend and/or filter semantically close or semantically disparate candidate files from a query from a user that includes the data file.

Abstract: A system for analyzing screenshots can include a computing device including a processor coupled to a memory and a display screen configured to display content. The system can include an application stored on the memory and executable by the processor. The application can include a screenshot receiver configured to access, from storage to which a screenshot of the content displayed on the display screen captured using a screenshot function of the computing device is stored, the screenshot including an image and a predetermined marker. The application can include a marker detector configured to detect the predetermined marker included in the screenshot. The application can include a link identifier configured to identify, using the predetermined marker, a link to a resource mapped to the image included in the screenshot, the resource accessible by the computing device via the link.

Abstract: An automatic analyzer includes an input reception unit that receives an operation to specify a search condition for sample information, an operation to execute search of the sample information, and an operation to specify whether only pieces of the sample information of samples under measurement among all samples are set as search targets; and a display control unit that, when the operation to execute search of the sample information is performed in a state in which only pieces of the sample information of samples under measurement are specified to be search targets, sets only pieces of the sample information of samples under measurement among the sample information of all samples as search targets based on the progress information, extracts pieces of the sample information of samples matching with a specified search condition, and causes the extracted pieces of the sample information to be displayed in a list form.

Abstract: A method for execution by a dispersed storage and task (DST) processing unit includes obtaining audit records for an audit object and determining when the audit object is complete. When the audit object is complete, aggregating the audit records of the audit object within the audit object by generating the audit object to include the audit records; generating identifier (ID) information and generating integrity information. Fields of the audit object are populated with the audit records, the ID information, and the integrity information and a name of the audit object is determined for storage of the audit object and the name of the audit object in a dispersed storage network (DSN).

Abstract: A method comprising: receiving a plurality of pairs of queries associated with a database, wherein the queries in each pair in the plurality of pairs of queries have an identical FROM clause; at a training stage, training a machine learning model on a training set comprising: (i) the plurality of pairs of queries, and (ii) labels associated with containment rates between each of the pairs of queries over the database; and at an inference stage, applying the trained machine learning model to a pair of target queries, to estimate containment rates between the target pair of queries over the database.

Abstract: One aspect of the disclosure relates to providing players pricing of in-game virtual items associated with their experience and their progress in the game. For example, the first player and the second player may have similar triggers, similar in-game events, and actually have different prices and currencies inside each story depending on each player's progress in the game, thus creating an individualized experience that may alter from player to player.

Abstract: A global filter allows data filtering using attributes across multiple Analysis Tools (ATs), by broadcasting semantic filter context objects. Upon selecting object attribute values, the filter context object is created with attribute names and values. A processing engine resolves the filter context object to a data object, and then subsequently to target data. A lateral filter finds related entities in a relational database, without having to maintain and/or duplicate all of the data into a graph database. The processing engine resolves lateral filters using an entity graph path calculation conducted in conjunction with the generation of a bootstrapped graph structure. That graph structure is constructed (bootstrapped) utilizing available database schematic information—e.g., pre-calculated (key) relations and metadata read from the relational database. From that information, relationships in the bootstrapped graph structure are determined.

Abstract: Methods, systems, apparatuses, and computer program products are provided for generating an instruction set for an evaluation engine. An arithmetic expression that combines multiple columns of data (e.g., a first column of data, a second column of data, etc.) is received. Instructions may be generated, that, when executed by an integrated-circuit-based processor, cause the integrated-circuit-based processor to evaluate the arithmetic expression. In examples, a set of instructions may be generated for each column of data represented in the arithmetic expression. For instance, the instructions may comprise a first set of instructions associated with the first column of data, a second set of instructions associated with the second column of data, and so on. The instructions may specify one or more parameters for operations associated with each column of data, such as operations to load data from a buffer, store data into a buffer, arithmetic operations to perform on data, etc.

Abstract: An augmented reality (AR) providing method for recognizing a context using a neural network includes acquiring, by processing circuitry, a video; analyzing, by the processing circuitry, the video and rendering the video to arrange a virtual object on a plane included in the video; determining whether a scene change is present in a current frame by comparing the current frame included in the video with a previous frame; determining a context recognition processing status for the video based on the determining of whether the scene change is present in the current frame; and in response to determining that the context recognition processing status is true, analyzing at least one of the video or a sensing value received from a sensor using the neural network and calculating at least one piece of context information, and generating additional content to which the context information is applied and providing the additional content.

Abstract: An illustrative “open archive” architecture provides streamlined access to production data, which originally was stored on a NAS device but which is archived to secondary storage to free up NAS space. An open-archive server coordinates with an open-archive layer on the NAS device. The open-archive server identifies data sets on the NAS that meet archiving criteria, which are then automatically moved to an open archive in secondary storage. The open archive layer intercepts data-access calls coming into the NAS device, and reports the intercepted calls to the open-archive server. If the open-archive server determines that the data referenced in an intercepted call is in the open archive, the server initiates a restore job that recovers the data from secondary storage and stores it back on the NAS device. The intercepted call may now be served from the NAS. These operations occur automatically and without data agents for the NAS-based data.

Abstract: The disclosed embodiments provide a system for processing data. During operation, the system identifies content shared within an online system and interactions between users of the online system and the content that match behavioral criteria and content-based criteria in a set of rules. Next, the system aggregates the content and the interactions into trends in the content and the interactions, predictions associated with the content and the interactions, and recommendations for moderating the content and the interactions. The system then outputs representations of the trends, the predictions, and the recommendations in a user interface. Finally, the system receives, via the user interface, an action to be performed on a subset of the content and updates sharing of the content within the online system based on the action.

Abstract: The present disclosure provides method and apparatus for processing a message. A statement sentence message and a message processing parameter associated with a user's session message are obtained. One or more first statement sentence nodes that are semantic-matched with the statement sentence message are determined in the knowledge map. One or more second statement sentence nodes corresponding to the message processing parameters are obtained from the knowledge map, based on the node relationship properties of the first statement sentence nodes. A response is generated based at least in part on statement sentences of the one or more second statement sentence nodes. The generated response is provided to the user.

Abstract: A system for generating adaptive notifications including processors and memories storing instructions that configure the processors to perform operations. The operations may include generating an item list for notifications to a client ID, selecting a first item from the item list for notification, the first item being selected having a highest priority score and a lowest lifecycle timer, encoding a first landing page by encoding a product code and a tracking cookie in a URL configured for redirection to a first webpage by a redirect server, and transmitting a first notification with a payload including the first landing page and a first message. The operations may also include after transmitting the first notification, modifying priority scores and updating lifecycle timers based on determining whether the client device visited the first landing page. Further, the operations may also include selecting a second item from the item list for notification.

Abstract: System and method of generating an executable action item in response to natural language dialogue are disclosed herein. A computing system receives a dialogue message from a remote client device of a customer associated with an organization, the dialogue message comprising an utterance indicative of an implied goal. A natural language processor of the computing system parses the dialogue message to identify one or more components contained in the utterance. The planning module of the computing system identifies the implied goal. The computing system generates a plan within a defined solution space. The computing system generates a verification message to the user to confirm the plan. The computing system transmits the verification message to the remote client device of the customer. The computing system updates an event queue with instructions to execute the action item according to the generated plan upon receiving a confirmation message from the remote client device.

Abstract: A system and method are described for managing attributes in an IoT system.

Abstract: A computer-implemented method and a smart-learning and knowledge retrieval system (SLKRS) are provided for imparting adaptive and personalized e-learning based on continually artificially learned unique characteristics of a knowledge seeker. The SLKRS ingests data in multiple formats from multiple sources, merges the data into a knowledge base based on computed strengths of terms in the sources, and assimilates the merged data to generate experiences. In response to a query received from the knowledge seeker, the SLKRS retrieves and sends in an immersive format one of the generated experiences or an experience created based on an artificially intelligent understanding of the received query. The SLKRS receives feedback from the knowledge seeker and computes a score based on the feedback and the query to artificially learn unique characteristics of the knowledge seeker. The SLKRS generates interventions and improved experiences for the knowledge seeker based on the computed score.

Abstract: In an information processing apparatus having a platform supporting estimation processing using a model that has learned a relationship between data associated with user word and action and information to be used for search processing, in a case where the estimation processing is executed by using input data based on user input performed before usage of an application and the model, and an output estimation result is recorded, the application acquires the estimation result from the platform, executes search processing using information included in the estimation result, and provides the information using the result.

Abstract: A method of identifying one or more objects in an image file that includes receiving an image file at a first classifier, wherein the image file comprises image metadata that identifies characteristics of the image file. The method includes determining whether the first classifier is trained to identify at least one object in the image file based on a first set of training metadata associated with the first classifier, wherein the first set of training metadata defines a set of characteristics for which the first classifier is trained to identify the at least one object. The method includes identifying the one or more objects in the image file using the first classifier responsive to determining that the first classifier is trained to identify the at least one object.

Abstract: A query change system includes: a processor configured to correct, in a case where a first query image inputted by a user includes a contradicting part that contradicts a first condition related to a search target, the contradicting part of the first query image in accordance with the first condition to generate a second query image.

Abstract: A system and method for generating encryption keys on multiple devices, without transferring the keys. At least one sender memristor is set using at least one sender setting value. At least one sender reading value is applied to the at least one sender memristor to generate at least one sender output value. A string of characters is determined from the at least one output value based on a sender table. Data is encrypted with the string of characters. The encrypted data is transmitted to a receiver through a first channel. The at least one sender setting value or the at least one sender reading value or both is transmitted to the receiver through a second channel different from the first channel. The at least one sender setting value or the at least one sender reading value or both is applied to at least one receiver memristor to generate at least one receiver output value. A receiver table is used to determine the string of characters from the at least one receiver output value.

Abstract: A medical expense document management system including a document management server is disclosed. The server is configured to receive from a user client device an image of a medical document, convert the image into a plurality of text elements using OCR, and determine a source of the document. The server is also configured to retrieve data detectors from a database, each data detector associated with a data type anticipated to be in the document, each detector having at least one identifier, at least one direction describing a potential relative direction of a text element having a label associated with the detector, and at least one validation criteria. The server is also configured to identify a potential descriptor by comparing each text element with the at least one identifier, determine if the text element pointed to by one of the directions meets the validation criteria, and store the validated text element.