Amazon Technologies, Inc.

James Claiborne Moore

Gaurav Saxena

Kiran Kumar Chinta

Abhishek Sharma

Todd Jeffrey Green

Yanzhu Ji

Dmytro Bogatov

Ameya Ashok Limaye

Siyong Liang

Yasser Gonzalez Fernandez

Majid Laali

Techniques for priority-based masking policy selection in a database environment are described. Masking policies are defined and attached to columns of relational data for particular users or roles. The attachment of a masking policy to a column includes a user-specified priority value. When multiple policies could apply to a particular query, the conflict can be easily resolved and understood by use of the priority values, for example, by selecting a candidate policy having a highest priority value.

Priority-based masking policy selection in a database environment

Techniques for dynamic database redaction using protected encryption secret material are described. A masking policy is defined that includes a reference to a secret material stored by a secrets manager service. The masking policy further identifies a pseudonymous redaction function that utilizes a cryptographic function requiring such a secret material. The secrets manager service is configured to grant access to the secret material by an entity of the database service that executes queries, such as a leader node of a cluster. For a particular query, the cluster obtains the secret material from the secrets manager service in a secure manner, uses the secret material for applying the cryptographic function to values for redaction purposes, and deletes any copies of secret material thereafter.

Dynamic database redaction using protected secret material

A phonetic search system may pass phonetic information from an automatic speech recognition (ASR) system to a natural language understanding (NLU) system for the latter to leverage when performing entity resolution in the presence of ambiguous interpretations. The ASR system may include an acoustic model and a language model. The acoustic model can process audio data to generate hypotheses that can be mapped to acoustic data; i.e., one or more acoustic units such as phonemes. The language model can process the acoustic units to generate text data representing possible transcriptions of the audio data. ASR/NLU systems may have difficulty interpreting speech when confronted with, for example, homographs, which are words that are spelled the same, but have different meanings. When uncertainty in the final transcription is high, the system can leverage the acoustic data to improve the accuracy of entity resolution.

Entity resolution using acoustic data

Growing community of inventors

Boston, MA, United States of America

James Claiborne Moore