A method and a system for generating an interpretable embedding that corresponds to a sequence of events is provided. The method includes: receiving information that corresponds to a sequence of events that respectively correspond to interactions between a customer and an organization; determining, for each respective event, a respective product associated with the organization and a respective channel via which the event has occurred; assigning a respective sentiment to each event; computing a respective weight for each event; aggregating the computed weights with respect to the products and the channels; and using the aggregated weights to generate the interpretable embedding for the customer. The interpretable embedding is then usable for generating targeted offers to the customer, handling complaints, and preventing subsequent complaints.

The technology described herein determines whether a candidate text is in a requested class by using a generative model that may not be trained on the requested class. The present technology may use of a model trained primarily in an unsupervised mode, without requiring a large number of manual user-input examples of a label class. The may produce a semantically rich positive example of label text from a candidate text and label. Likewise, the technology may produce from the candidate text and the label a semantically rich negative example of label text. The labeling service makes use of a generative model to produce a generative result, which estimates the likelihood that the label properly applies to the candidate text. In another aspect, the technology is directed toward a method for obtaining a semantically rich example that is similar to a candidate text.

Systems and methods for data warehouse management for medical information is provided. The system receives a set of medical record documents. These documents include evidence for one or more findings which are identified using natural language processing of evidence. The data set is probabilistically transformed into a structured data set (often as a table). This structured data set includes embedded links that reference the source document where the evidence is located. If the finding is supported by multiple articles of evidence, the link will direct the user to the source document with the highest confidence ranking. Evidence in the source document is highlighted and may include associated annotations. Evidence, findings and annotations may be updated by the user.

Method, system, and programs for determining a keyword from user agent strings are disclosed. In one example, a plurality of user agent strings is received. The plurality of user agent strings is grouped into one or more clusters. The one or more clusters comprise a first cluster that includes two or more user agent strings. The two or more user agent strings in the first cluster are compared. Based on the comparing, a keyword is determined from the first cluster. The keyword represents a type of user agent information.

In response to running at least one testing phrase on a previously trained text classifier and identifying a separate predicted classification label based on a score calculated for each respective at least one testing phrase, a text classifier decomposes extracted features summed in the score into word-level scores for each word in the at least one testing phrase. The text classifier assigns a separate heatmap value to each of the word-level scores, each respective separate heatmap value reflecting a weight of each word-level score. The text classifier outputs the separate predicted classification label and each separate heatmap value reflecting the weight of each word-level score for defining a heatmap identifying the contribution of each word in the at least one testing phrase to the separate predicted classification label for facilitating client evaluation of text classification anomalies.

The present disclosure provides systems and methods that include or otherwise leverage use of a multiscale quantization model that is configured to provide a quantized dataset. In particular, the multiscale quantization model can receive and perform vector quantization of a first dataset. The multiscale quantization model can generate a residual dataset based at least in part on a result of the vector quantization. The multiscale quantization model can apply a rotation matrix to the residual dataset to generate a rotated residual dataset that includes a plurality of rotated residuals. The multiscale quantization model can perform reparameterization of each rotated residual in the rotated residual dataset into a direction component and a scale component. The multiscale quantization model can perform product quantization of the direction components of the plurality of rotated residuals, and perform scalar quantization of the scale components of the plurality of rotated residuals.

At least one table included in first content may be determined. The at least one table includes a first plurality of rows and a first plurality of columns. It may be determined that a first term indicative of a personal name is included in a first row of the first plurality of rows and a first column of the first plurality of columns. A second row of the first plurality of rows that includes at least a first personal name in the first column and a first item of personal identifying information in a second column of the first plurality of columns may be identified. First data indicative of the first personal name and the first item of personal identifying information may be extracted. The first data may be added to a first profile associated with the first personal name.

Systems and methods for generating customized annotations of a medical record are provided. The system receives a medical record and processes it using a predictive model to identify evidence of a finding. The system then determines whether to have a recall enhancement or validation of a specific finding. Recall enhancement is used to tune or develop the predictive model, while validation is used to rapidly validate the evidence. The source document is provided to the user and feedback is requested. When asking for validation, the system also highlights the evidence already identified and requests the user to indicate if the evidence is valid for a particular finding. If recall enhancement is utilized, the source document is provided and the user is asked to find evidence in the document for a particular finding. The user may then highlight the evidence that supports the finding. The user may also annotate the evidence using free form text.

A contextual index compendium that includes contextual index item generation rules that define document index entry generation transforms usable to transform text of the documents into embedded document index entries of document indexes within the documents is obtained by a processor. Using the document index entry generation transforms defined within the contextual index item generation rules in association with a document that includes embedded document index entries that are both embedded at locations of associated text distributed throughout the document and added as part of a document index within the document, new text of the document is programmatically transformed into at least one new document index entry in response to determining that at least one portion of the new text includes candidate text that is not already indexed within the existing embedded document index entries and the document index within the document.

This application relates to a systems and methods for trending issue identification in text streams. In one embodiment, a method for improving resolution of a trending issue identified in a set of text streams includes presenting a user interface of an application that is being executed by a computing device. The method also includes receiving a notification including the trending issue that has been identified in the set of text streams based at least in part on textual analysis performed on the set of text streams, and presenting the trending issue on the user interface of the application to enable an action to be performed to resolve the trending issue.