System and method for phonetic hashing and named entity linking from output of speech recognition
AssigneeFresh Consulting, Inc
InventorsRobert HILD, Sean Alan McKAY, Eli RODRIGUEZ
Life of this patent
Estimating the term…
Abstract
A system and method for named entity linking from the output of speech-to-text systems by using an approximate string matching that normalizes common sounds, removes ambiguities, removes silent consonants, and accounts for speech slurring for long names. Additionally, the system and method for named entity linking from the output of speech-to-text systems employs a hierarchical matching system that performs multiple attempts using various mechanisms for resolving the name, starting with a very strict mechanism, and proceeding sequentially through less strict mechanisms.
Claims24
A method for named entity linking from the output of speech-to-text systems by reducing names to common sound codes, the method comprising:
- performing phonetic hashing and disambiguation that normalize common sounds and cause similar sounds to generate a same output;
- checking for partial matches by including a creolizing and comparing disambiguation process;
- removing ending letters from longer names, thereby accounting for speech slurring that occurs with longer names;
- and
- performing multiple attempts to resolve non-exact matches using a hierarchical matching system, starting with a strict mechanism, and proceeding sequentially through less strict mechanisms, wherein the hierarchical matching system provides a high resistance to false positives.
A system for named entity linking from the output of speech-to-text systems by reducing names to common sound codes, the system comprising:
- one or more processors;
- and
- a memory device storing a set of instructions that when executed by the one or more processors, causes the one or more processors to:
- perform phonetic hashing and disambiguation that normalize common sounds and cause similar sounds to generate a same output;
- checking for partial matches by including a creolizing and comparing disambiguation process;
- remove ending letters from longer names, thereby accounting for speech slurring that occurs with longer names;
- and
- perform multiple attempts to resolve non-exact matches using a hierarchical matching system, starting with a strict mechanism, and proceeding sequentially through less strict mechanisms, wherein the hierarchical matching system provides a high resistance to false positives.
A system for identifying words or commands from audio information, the system comprising:
- one or more processors;
- and
- a memory device storing a set of instructions that, when executed by the one or more processors, causes the one or more processors to:
- perform phonetic hashing and disambiguation that normalize common sounds and cause similar sounds to generate a same output;
- checking for partial matches by including a creolizing and comparing disambiguation process;
- and
- remove ending letters from longer words, thereby accounting for speech slurring that occurs with longer words.
Dependencies are inferred from the claim text.