International Business Machines Corporation

American Express Travel Related Services Company, Inc.

Salil Joshi

Prasad Manikarao Deshpande

Shourya Roy

Atreyee Dey

Sameep Mehta

Mukesh Kumar Mohania

Sriram Kolinjivadi Padmanabhan

Sanjeev K Gupta

Rajeev Gupta

Manish Kesarwani

Arpan Somani

Scott Schumacher

Mohit Jain

Karin Murthy

Hima Prasad Karanam

Songhua Xing

Dawn Thomas

Bharath Venkatesh

Hima P Karanan

Yue Jiao

Bruhathi H Sundarmurthy

At least some embodiments are directed to an entity matching detection system. The entity matching detection system includes a latent similarity identification machine learning model that receives one or more data records and generates a final similarity score indicative of a latent similarity between the one or more data records and a second data record. The entity matching detection system can identify lexical and semantic similarities between attribute values and can analyze and compute similarity scores for direct-linked attribute values and cross-linked attribute values extracted from different data records.

Computer-based systems for data entity matching detection based on latent similarities in large datasets and methods of use thereof

Disclosed are various approaches for providing a natural language interface for searching databases. A natural language query is parsed to identify a plurality of tokens. At least one operator is identified with a machine-learning model based at least in part on the plurality of tokens. Next, at least one attribute and at least one respective attribute value are identified with a machine-learning model based at least in part on the plurality of tokens. Then, at least one constraint is identified with a machine-learning model based at least in part on the plurality of tokens. Finally, a machine language query is generated based at least in part on the at least one operator, the constraint(s), the attribute(s), and the respective attribute value(s).

Natural language interface to databases

Methods, systems, and computer program products for tracking missing data using provenance traces and data simulation are provided herein. A computer-implemented method includes generating, for each of multiple stages in a data curation sequence, a machine learning model of the data curation sequence, wherein the model is based on historical input records within the data curation sequence, historical output records within the data curation sequence, and provenance data within the data curation sequence; creating a simulated output record based on a detected anomaly corresponding to the data curation sequence; predicting the content of absent input records that precede the simulated output record in the data curation sequence and provenance data corresponding to the simulated output record; and outputting, to a user, in response to a query pertaining to the detected anomaly, the predicted input records and information relating the predicted input records to the detected anomaly.

Tracking missing data using provenance traces and data simulation

Methods, systems and computer program products for root cause analysis using provenance data are provided herein. A computer-implemented method comprises computing a plurality of provenance paths for at least one of a plurality of data elements in a curation flow and a plurality of groups of data elements in the curation flow, analyzing the computed provenance paths to determine one or more errors in the curation flow, and outputting the one or more errors in the curation flow to at least one user. The analyzing comprises at least one of identifying which of the computed provenance paths are partial provenance paths, and identifying one or more output records associated with the curation flow, wherein the one or more output records comprise incorrectly curated data, and identifying the computed provenance paths that respectively correspond to the one or more output records comprising the incorrectly curated data.

System and method for root cause analysis in large scale data curation flows using provenance

Entity mappings that produce matching entities for a first data asset having attributes and a second data asset having attributes are generated by: generating entity mappings that produce matching entities for a first data asset having attributes with attribute values and a second data asset having attributes with attribute values by: matching the attribute values of the attributes of the first data asset with the attribute values of the attributes of the second data asset, using the matching attribute values to generate matching attribute pairs, and using the matching attribute pairs to identify entity mappings; computing an entity mapping score for each of the entity mappings based on a combination of factors; ranking the entity mappings based on each entity mapping score; and using some of the ranked entity mappings to determine whether a same real-world entity is described by the first data asset and the second data asset.

Identifying entity mappings across data assets

Methods and arrangements for discovering entity types for a set of records. A set of records is input, with each record comprising attributes with associated attribute values. The records are grouped into candidate entity types in view of at least one of: the attribute values of the records, at least one domain ontology and at least one dimension hierarchy. An interestingness measure of each candidate entity type is calculated, via estimating interestingness based on at least one factor selected from the group consisting of: a correlation between attribute values of records, a number of attributes, a log of queries issued to a server, and an average group size for candidate entity types. At least one candidate entity type is validated based on the calculated interestingness measures. Other variants and embodiments are broadly contemplated herein.

Discovery of related entities in a master data management system

Embodiments of the present invention disclose a method, computer program product, and system for identifying matching products relative to a reference product. A reference product is identified from a received product query and a query is generated based on the reference product. A generated query comprises of an ontology, at least one word appearing in a title of the reference product, and a set of key words appearing in social media data associated with the reference product. A database is searched using the generated query to find matching product sets and the results are returned and filtered. Results are filtered by calculating a relationship score between the reference product and one or more matching products in the set of matching products, and/or by filtering a subset of the set of matching products based on a customer profile. The filtered subset of results are communicated to a recipient.

Product recommendations over multiple stores

Methods and arrangements for facilitating a display of search results. A search query is received from a user having a predetermined access level. A search is executed based on the search query and, based on the executed search, initial search results are produced. Sensitive information is redacted from the initial search results, based on the predetermined access level. The redacted search results are filtered and re-ordered to forestall an inference of the redacted sensitive information. The filtered and re-ordered search results are displayed to a user. Other variants and embodiments are broadly contemplated herein.

Growing community of inventors

Bengaluru, India

Salil Joshi