Amazon Technologies, Inc.

Stephen Michael Ash

Bing Xiang

Zhiguo Wang

Timothy Jones

Andrew Borthwick

Patrick Ng

Sudipta Sengupta

Ramesh M Nallapati

Sharanabasappa Parashuram Revadigar

Gregory David Adams

Hanbo Li

Rishav Chakravarti

Jiarong Jiang

Anurag Windlass Gupta

Mehul A Shah

Mukul Vijay Karnik

Mehul Y Shah

Prajakta Datta Damle

Douglas Scott Cerny

Juliana Saussy

Yung Haw Wang

Abhishek Sharma

Nanyan Li

Jun Wang

Stephanie Weeks

Shehzad Qureshi

Pedrito Uriah Maynard-Zhang

Adam Lawrence Joseph Heinermann

Deepak Shantha Murthy

Alaykumar Navinchandra Desai

Anuj Chauhan

Sergei Dobroshinsky

Donald Harold Rivers Weidner

Chethan Kommaranahalli Rudramuni

Henghui Zhu

Russell Reas

William Michael Siler

Rajkumar Haridoss

Kushal Yelamali

Rajesh Patel

Amjad Al-Rikabi

Wuwei Lan

Sheng Zhang

Robert Anthony Barton, Jr

Techniques for handling natural language query processing are described. In some examples, entities are recognized during an entity recognition phase and then relations between those entities are determined. Those relations are fed to an entity linker to help the linker link candidate to columns and/or a intent representation generator to help parse multiple values and column pairs of a natural language query.

Natural language query processing

Techniques for handling natural language query processing are described. In some examples, a sequence-to-sequence model is used to handle a natural language query. Post-processing of a result of the sequence-to-sequence model utilizes fine-grained information from an entity linker. In some examples, the sequence-to-sequence model and aspects of a natural language query pipeline are used to handle a natural language query.

This disclosure describes a natural language question (NLQ) query service within a service provider network that provides row level security (RLS) for autocomplete during entry of NLQs and fuzzy matching in NLQ answering. The rules take the form of per-user predicates such as Tim can only see rows with region=US. In configurations a complex extraction and preprocessing pipeline to extract distinct combinations of values against RLS predicate 'rule keys' is used. Those distinct values are indexed along with grouped rule keys to enable pushing down predicates at auto-complete time. This enables pushing part of RLS rule handling to ingestion time of a dataset rather than handling all RLS rule handling at query time, enabling meeting of latency goals. In configurations, a single logical document of unique cell values is split into multiple documents with a subset of rule keys to handle scalability limits.

Row level security in natural language question answering

Interactive assistances for executing natural language queries to data sets may be performed. A natural language query may be received. Candidate entity linkages may be determined between an entity recognized in the natural language query and columns in data sets. The candidate linkages may be ranked according to confidence scores which may be evaluated to detect ambiguity for an entity linkage. Candidate entity linkages may be provided to a user via an interface to select an entity linkage to use as part of completing the natural language query.

Interactive assistance for executing natural language queries to data sets

Specified performance attributes may be used to configure machine learning transformations for ETL jobs. Performance attributes for a machine learning pipeline that applies a model to as part of a transformation for an ETL job may be used to configure a parameter in a stage of the machine learning pipeline. The configured stage may then be used when training the model. The trained machine learning pipeline may then be applied as part of a transformation operation included in an ETL job performed by the ETL system.

Using specified performance attributes to configure machine learning pipepline stages for an ETL job

Multiple stage filtering may be implemented for natural language query processing pipelines. Natural language queries may be received at a natural language query processing system and processed through a query language processing pipeline. The query language processing pipeline may filter candidate linkages for a natural language query before performing further filtering of the candidate linkages in the natural language query processing pipeline as part of generating an intermediate representation used to execute the natural language query.

Multiple stage filtering for natural language query processing pipelines

Supervised partitioning is used to perform record matching. A request to identify matches between records is received. A graph representation that indicates similarities between the records is partitioned and an evaluation of the partitioning is performed according to a supervised machine learning technique to generate a confidence value in the partitioning. An indication of equivalent records according to the partitioning and the confidence value of the partitioning may be provided.

Supervised graph partitioning for record matching

Techniques for memory-efficient streaming count estimation for multisets are described. A method for memory-efficient streaming count estimation for multisets may include obtaining data from a plurality of data sources, and estimating a count for one or more attributes of the data using a telescoping count-min sketch (CMS) data structure, the telescoping CMS including at least a first table and a second table, wherein count values for the data are stored in a plurality of cells of the first table and when a cell of the first table is saturated, the count values for that cell are stored in a corresponding cell of the second table determined based at least on the cell of the first table.

Memory-efficient streaming count estimation for multisets

Using specified performance attributes to configure machine learning pipeline stages for an ETL job

A data records service is configured to receive original data records and, in parallel, generate a transliterated version of the original data record into a phonetic based language. Individual fields of data records can be transliterated by identifying a primary language, generating language specific tokens for individual text portions, and transliterating the token. The records processing service can then execute matching models on both original data records and transliterated data records to detect matching data records.

Transliteration of data records for improved data matching

Techniques for scaling record linkage via elimination of highly overlapped blocks are described. A method for scaling record linkage via elimination of highly overlapped blocks includes identifying a first plurality of blocks based at least on a plurality of records stored in a storage service of a provider network, identifying a plurality of sets of matching blocks from the first plurality of blocks, deleting the plurality of sets of matching blocks except for a first block from each set from the plurality of sets of matching blocks, and iteratively performing dynamic blocking based at least on the first block to generate subsequent pluralities of blocks until the subsequent pluralities of blocks are below a threshold size.

Scaling record linkage via elimination of highly overlapped blocks

Techniques for Scalable parallel elimination of approximately subsumed sets are described. A method for Scalable parallel elimination of approximately subsumed sets includes identifying a first plurality of blocks based at least on a plurality of records stored in a storage service of a provider network, determining a plurality of subsumption relationships between blocks from the first plurality of blocks, retaining a first subset of the first plurality of blocks and demoting a second subset of the first plurality of blocks based at least on the plurality of subsumption relationships, and iteratively performing dynamic blocking based at least on the first subset of the plurality of matching blocks and the second subset of the plurality of matching blocks to generate a subsequent pluralities of blocks.

Scalable parallel elimination of approximately subsumed sets

Growing community of inventors

Seattle, WA, United States of America

Stephen Michael Ash