Abstract: A method includes processing a set of query texts to identify a set of expressions, where each expression references a set of columns of datetime data in a datastore. The method also includes training a statistical model to determine a distribution of the datetime data for each column that was identified. The method further includes processing the set of expressions to generate a directed graph including more than one nodes and a plurality of edges, where each node represents one of the columns or a transformation applied by one of the expressions to one of the columns. The method additionally includes generating a weight for edges of the directed graph according to a distribution of the datetime data in the columns and a usage index of a corresponding expression.

Abstract: A method may including obtaining a voice transcript corpus and a chat transcript corpus, extracting voice transcript sentences from the voice transcript corpus and chat transcript sentences from the chat transcript corpus, encoding, by a series of neural network layers, the voice transcript sentences to generate voice sentence vectors, encoding, by the series of neural network layers, the chat transcript sentences to generate chat sentence vectors, determining, for each voice sentence vector, a matching chat sentence vector to obtain matching voice-chat vector pairs, and adding, to a parallel corpus, matching voice-chat sentence pairs using the matching voice-chat vector pairs. Each of the matching voice-chat sentence pairs may include a voice transcript sentence and a matching chat transcript sentence. The method may further include training a disfluency remover model using the parallel corpus.

Abstract: A processor may receive an image and identify a plurality of characters in the image using a machine learning (ML) model. The processor may generate at least one word-level bounding box indicating one or more words including at least a subset of the plurality of characters and/or may generate at least one field-level bounding box indicating at least one field including at least a subset of the one or more words. The processor may overlay the at least one word-level bounding box and the at least one field-level bounding box on the image to form a masked image including a plurality of optically-recognized characters and one or more predicted fields for at least a subset of the plurality of optically-recognized characters.

Abstract: Certain aspects of the present disclosure provide techniques for interacting with a graph database structure. In one embodiment, a method includes receiving, at an application, information regarding a first entity; transmitting, to a graph database, a query regarding the first entity; receiving, at the application, query results based on one or more relationships between the first entity and other entities in the graph database; making, by the application, an inference based on the query results; modifying, by the application, a user interface of the application based on the inference by displaying at least one user interface element suggesting a selection of an application option; and receiving, by the application, a user selection of the suggested application option.

Abstract: Certain aspects of the present disclosure provide techniques for optimizing results generated by functions executed using a rule-based knowledge graph. The method generally includes generating a neural network based on a knowledge graph and inputs for performing a function using the knowledge graph. Inputs for the function are received and used to generate a result of the function. A request to optimize the generated result of the function is received. A loss function is generated for the neural network. Generally, the loss function identifies a desired optimization for the function. Values of parameters in the neural network are adjusted to optimize the generated result based on the generated loss function, and the adjusted values of the parameters in the neural network are output in response to the request to optimize the generated result of the function.

Abstract: Systems and methods for mapping an electronic document to a particular transaction category are disclosed. An example method may be performed by one or more processors of a categorization system and include receiving, from a user via an interface, an electronic document associated with a transaction between the user and a particular establishment, receiving, from the user via the interface, an image of the particular establishment, identifying in the image, using an image recognition engine, at least one of a sign or a symbol representative of the particular establishment, extracting, using an analytics module, location information from at least one of the image or a mobile device, determining, using the analytics module, a name of the particular establishment based on at least one of the location information or the at least one sign or symbol, and mapping the electronic document to a particular transaction category based on the determined name.

Abstract: A method including generating a captured facial object and a captured pose from a captured image. The method also includes obtaining a base facial object and a base pose from a base image. The method also includes generating base pose angles using the captured pose, and captured pose angles using the captured pose. The method also includes obtaining selected base images using the base pose angles and the base facial object. The method also includes generating selected captured images using the captured pose angles and the captured facial object. The method also includes comparing the selected base images to the selected captured images to establish a comparison. The method also includes outputting a match output using the comparison.

Abstract: Embodiments disclosed herein may extract trending topics from phone call transcripts or any type of text data. The phone call transcripts may be collected for a time period and the time period may be divided into time spans. For each time span having more than a threshold number of phone call transcripts, n-grams from the phone call transcripts may be extracted. The extracted n-grams may be contextually clustered by converting the n-grams into their embedding vectors, reducing the dimensionality of the embedding vectors, and clustering similar reduced dimensionality embedding vectors. Normalized occurrences of one or more clusters may be generated. The recent mean of the number of occurrences of the normalized clusters may be compared with the historical mean and offset by historical standard deviation to generate a modified Z-score. N-grams corresponding to the clusters with high Z-scores may be identified as trending topics.

Abstract: A method of score prediction uses hierarchical attention. Word features, positioning features, participant embedding features, and metadata are extracted from a transcript of a conversation. A word encoder vector is formed by multiplying weights of a word encoder layer to one or more word features. A sentence vector is formed by multiplying weights of a word attention layer to word encoder vectors. An utterance encoder vector is formed by multiplying weights of an utterance encoder layer to the sentence vector. A conversation vector is formed by multiplying weights of an utterance attention layer to utterance encoder vectors. The utterance encoder vector is combined with one or more positioning features and one or more participant embedding features. A predicted net promoter score is generated by multiplying weights of an output layer to the conversation vector combined with the metadata. The predicted net promoter score is presented in a list of conversations.

Abstract: A method implements influencer segmentation detection. The method includes selecting transaction data for a time window and processing the transaction data for the time window to generate a graph for the time window. The method further includes extracting, from the graph, a feature set for a node of the graph for the time window and processing the feature set to generate a predicted rank for the node for a subsequent time window using a machine learning model. The method further includes selecting, using the predicted rank, an entity identifier corresponding to the node and presenting the entity identifier.

Abstract: Systems and methods for training machine learning models are disclosed. An example method includes receiving historical event timing data including event data for a first portion including events from a first time period, and a second portion comprising events from a second time period not including the first time period, predicting, based on the first portion of the historical event timing data, a first plurality of predicted events, the first plurality of predicted events corresponding to the second time period, determining a first subset of predicted events to be accurate predictions based at least in part on comparing the first plurality of predicted events to the historical events occurring within the second time period, generating training data based at least in part on the first subset of the first plurality of predicted events, and training the machine learning model based at least in part on the training data.

Abstract: A method services competing updates from multiple servicing instances. An update message is received by a class of a plurality of classes that service an object. The update message includes an update to the object and an expected version for the object. A version match is determined between the expected version and a current version of the object. When the version match is successful, the update is applied to the object. A response message including the current version of the object is generated. The response message is transmitted. When the update is successfully applied, the response message is transmitted as a success message. When the update is not successfully applied, the response message is transmitted as a failure message.

Abstract: Aspects of the present disclosure provide techniques for automated health scoring through meta-learning. Embodiments include retrieving text data related to an entity that was provided by a user and providing one or more first inputs to a first machine learning model based on a subset of the text data. Embodiments include determining, based on an output from the first machine learning model, whether the text data includes an address. Embodiments include determining that the text data includes a name and determining, based on the address and the name, that one or more text results from one or more data sources relate to the entity. Embodiments include providing one or more second inputs to a second machine learning model based on the one or more text results and determining, based on an output from the second machine learning model, a health score for the entity.

Abstract: The present disclosure provides techniques for data retrieval using machine learning. One example method includes receiving a plurality of training episodes associated with different environments, wherein each training episode of the plurality of training episodes includes a sequence of states, computing, based on the plurality of training episodes, total counts of a plurality of values in the states, initializing, for each state of the sequence of states in each training episode of the plurality of training episodes, a reward based on the total counts of the plurality of values, and training a reinforcement learning agent using the rewards.

Abstract: An access graph is constructed based on access policy data from user accounts, data lake buckets, and/or access policy statements from any other location. Access logs are analyzed to determine actual access to the data tables. For a given user role, an initial set of data tables that are actually accessed is generated forming the baseline of data tables for which access privileges are to be maintained. User roles that are similar to the given user role are identified and additional data tables accessed by the similar user roles are added to the initial set of data tables to generate a final set of data tables. Access privileges to the final set of data tables are maintained for the given user role, while access privileges to the remaining data tables may be revoked.

Abstract: A method including transcribing, into digital tokens, utterances from a conversation between an agent and a person. The method also includes embedding the digital tokens into an utterances tensor including sequences of the digital tokens. The method also includes obtaining a metadata tensor by encoding metadata related to the utterances into the metadata tensor. The method also includes executing a machine learning model which takes, as input, the utterances tensor and the metadata tensor, and which outputs a predicted source article predicted to be related to the utterances. The method also includes generating an interactive link to the predicted source article.

Abstract: Aspects of the present disclosure relate to watermarks and watermarking techniques for data streaming pipelines. Time stamp and offset timeline data is shared by computing instances along the pipeline to enable improved watermarking of the data stream through the pipeline. The improved watermarks enable better determination of completeness for the data stream and improve materialization of the results. The watermarking techniques can include periodically publishing watermark data by processing units of a vertex, fetching a merged watermark for a vertex by a vertex, and/or watching a data storage for the watermark data for events. Consensus algorithms can be used to maintain consensus among vertices for the watermark data.

Abstract: Disclosed are techniques for implementing an intelligent system with dynamic configurability. These techniques identifying a plurality of flow nodes for a software application and determine a dynamic flow for executions of the intelligent system with the plurality of flow nodes, one or more dynamic conditions, and one or more dynamic actions, without hard coded inter-dependency between two or more flow nodes of the plurality of flow nodes. The intelligent system is transformed into a dynamically configured intelligent system at least by performing a modification pertaining to one or more flow nodes in the dynamic flow, without affecting remaining flow nodes in the dynamic flow.

Abstract: A method is provided for authenticating a user. A request to access a resource is received from a user agent. A cookie associated with the request is identified. The cookie includes a first subset of data that was previously used to authenticate the user. The cookie is validated based on the first subset of the data. Responsive to validating the cookie, a second subset of the data is retrieved from server-side storage. A risk decision is generated based on the first subset and the second subset. When the risk decision meets a threshold, the user is authenticated without presenting an authentication challenge, and access to the resources permitted.

Abstract: A transcript of an audio conversation between multiple users (e.g., two users) is generated. The transcript is displayed in real time within a VR environment as the conversation takes place. A virtual selection tool is displayed within the VR environment to allow for a selection of different portions of the transcript. In addition, a virtual keyboard and or virtual panels with characters may be displayed and the virtual selection tool may be used to make selections from these displays as well. These selections are used to generate new text. The new text may form part of a user's notes of the conversation or an entry for a text field within the VR environment.