Various examples are directed to systems and methods for administering data model metadata for a plurality of data models. A metadata service may receive a first retrieval request from a requesting system. The first retrieval request may comprise an indication of a first local object identifier referencing a definition of a first local object from a first data model and an indication of a target data model. The metadata service may retrieve a first record from a metadata identifier registry, where the first record comprises an indication of the first local object identifier and an indication of a first global object identifier corresponding to the first local object identifier. The metadata service may determine a second local object identifier referencing a definition of a second local object identifier referencing a definition of a second local object in the target data model and return the second local object identifier.

Three-dimensional models (or avatars) may be defined based on imaging data captured from a customer. The avatars may be based on a virtual mannequin having one or more dimensions in common with the customer, a body template corresponding to the customer, or imaging data captured from the customer. The avatars are displayed on displays or in user interfaces and used for any purpose, such as to depict how clothing will appear or behave while being worn by a customer alone or with other clothing. Customers may drag-and-drop images of clothing onto the avatars. One or more of the avatars may be displayed on any display, such as a monitor or a virtual reality headset, which may depict the avatars in a static or dynamic mode. Images of avatars and clothing may be used to generate print catalogs depicting the appearance or behavior of the clothing while worn by the customer.

Systems and methods for managing data, such as metadata. In one exemplary method, metadata from files created by several different software applications are captured, and the captured metadata is searched. The type of information in metadata for one type of file differs from the type of information in metadata for another type of file. Other methods are described and data processing systems and machine readable media are also described.

Methods and systems for generating and using a conversation summary model. The method comprises receiving at least one training dataset. The at least one training dataset comprises data samples, each data sample comprising a text comprising text segments. The text is labelled with a conversation summary comprising any of the text segments which summarize the text. The at least one training dataset includes a dataset from a specific source. Using the at least one training dataset and the pre-trained model, the method further comprises generating the conversation summary model by fine-tuning the pre-trained model. The generated conversation summary model may be used to generate conversation summaries for chat conversations.

Methods and systems for generating and using a conversation summary model. The method comprises receiving at least one training dataset. The at least one training dataset comprises data samples, each data sample comprising a text comprising text segments. The text is labelled with a conversation summary comprising any of the text segments which summarize the text. The at least one training dataset includes a dataset from a specific source. Using the at least one training dataset and the pre-trained model, the method further comprises generating the conversation summary model by fine-tuning the pre-trained model. The generated conversation summary model may be used to generate conversation summaries for chat conversations.

Systems and method for lifelong tag learning. The system includes a computing device having a processor and a storage device storing computer executable code. The computer executable code is configured to: provide product descriptions and seed tags characterizing products; train a named-entity recognition (NER) model using the product descriptions and the seed tags; predict pseudo tags from the product descriptions using the NER model; calculate confidence scores of the pseudo tags; compare the confidence scores with a threshold, and define the pseudo tags as true tags when the confidence scores are greater than the threshold; add the true tags to the seed tags to obtain updated tags; and repeat the steps of training, predicting, calculating, comparing and adding using the product descriptions and the updated tags, so as to keep updating the updated tags.

Systems and method for lifelong tag learning. The system includes a computing device having a processor and a storage device storing computer executable code. The computer executable code is configured to: provide product descriptions and seed tags characterizing products; train a named-entity recognition (NER) model using the product descriptions and the seed tags; predict pseudo tags from the product descriptions using the NER model; calculate confidence scores of the pseudo tags; compare the confidence scores with a threshold, and define the pseudo tags as true tags when the confidence scores are greater than the threshold; add the true tags to the seed tags to obtain updated tags; and repeat the steps of training, predicting, calculating, comparing and adding using the product descriptions and the updated tags, so as to keep updating the updated tags.

Systems and methods identify and provide interesting facts about an entity. An example method includes selecting documents associated with at least one unique fact trigger, the documents being from a document repository. The method also includes generating entity-sentence pairs from the documents and, for a first entity of the entities represented by the entity-sentence pairs, clustering the entity-sentence pairs for the first entity using salient terms occurring in the sentence. The method also includes determining a representative sentence for each of the clusters and providing at least one of the representative sentences in response to a query that identifies the first entity. Another example method includes determining that a query relates to an entity in a knowledge base, determining that the entity has an associated unique fact list, and providing at least one of the unique facts in the list in response to the query.

Systems and methods identify and provide interesting facts about an entity. An example method includes selecting documents associated with at least one unique fact trigger, the documents being from a document repository. The method also includes generating entity-sentence pairs from the documents and, for a first entity of the entities represented by the entity-sentence pairs, clustering the entity-sentence pairs for the first entity using salient terms occurring in the sentence. The method also includes determining a representative sentence for each of the clusters and providing at least one of the representative sentences in response to a query that identifies the first entity. Another example method includes determining that a query relates to an entity in a knowledge base, determining that the entity has an associated unique fact list, and providing at least one of the unique facts in the list in response to the query.

Techniques for managing access to physical assets based on captured digital data and a database are provided. In one technique, one or more functions in an application that executes on a client device are locked. A smart badge that is associated with healthcare information is then received from a remote server system. In response to receiving the smart badge, the one or more functions are unlocked. After unlocking the one or more functions and in response to user input that selects a particular function of the one or more functions, a request and identification data that pertain to the particular function are transmitted over a computer network.