Back to results
US 11,769,012 B2Granted
Automated system and method to prioritize language model and ontology expansion and pruning
AssigneeVerint Americas Inc.
InventorsIan Roy Beaver, Christopher James Jeffs
Life of this patent
Estimating the term…
Abstract
A system and method for updating computerized language models is provided that automatically adds or deletes terms from the language model to capture trending events or products, while maximizing computer efficiencies by deleting terms that are no longer trending and use of knowledge bases, machine learning model training and evaluation corpora, analysis tools and databases.
Claims23
A non-transitory computer readable medium comprising instructions that, when executed by a processor of a processing system, cause the processor to perform a method of updating a language model for a language domain for an interactive virtual assistant, the method comprising:
- monitoring business related textual data from a plurality of platforms;
- collecting trending n-grams from the textual data over a sliding time window;
- comparing the n-grams to a vocabulary in a data model to identify terms comprising one of the trending n-grams existing in the data model;
- for a term comprising the one of the trending n-grams in the data model and appearing in a new context, wherein the new context is not associated with the trending n-gram in the data model, passing the term in the new context to a human for determination if the term should be added to a training example in the new context for retraining the data model;
- for a term comprising the one of the trending n-grams in the data model and not appearing in a new context, ignoring the term;
- for a term comprising the one of the trending n-grams and not in the data model, checking the term for frequency of use in the textual data and adding the term in context to the training example if the frequency of use has reached a predetermined threshold;
- recompiling the language model based on the term in context;
- and
- adopting the recompiled language model in an interactive virtual assistant.
6 more elements
A method of adding terms to a language model based on use, the method comprising:
- monitoring business related textual data from a plurality of platforms;
- collecting trending n-grams from the textual data over a sliding time window;
- comparing the n-grams to a vocabulary in a data model to identify terms comprising one of the trending n-grams existing in the data model;
- for a term comprising the one of the trending n-grams in the data model and appearing in a new context, wherein the new context is not associated with the trending n-gram in the data model, passing the term in the new context to a human for determination if the term should be added to a training example in the new context for retraining the data model;
- for a term comprising the one of the trending n-grams in the data model and not appearing in a new context, ignoring the term;
- for a term comprising the one of the trending n-grams and not in the data model, checking the term for frequency of use in the textual data and adding the term to the training example if the frequency of use has reached a predetermined threshold;
- recompiling the language model based on the term in context;
- and
- adopting the recompiled language model in an interactive virtual assistant.
6 more elements
A system for performing speech analytics, comprising:
- a memory comprising executable instructions;
- a processor configured to execute the executable instructions and cause the system to:
- monitor business related textual data from a plurality of platforms;
- collect trending n-grams from the textual data over a sliding time window;
- compare the n-grams to a vocabulary in a data model to identify terms comprising one of the trending n-grams existing in the data model;
- for a term comprising the one of the trending n-grams in the data model and appearing in a new context, wherein the new context is not associated with the trending n-gram in the data model, pass the term in the new context to a human for determination if the term should be added to a training example in the new context for retraining the data model;
- for a term comprising the one of the trending n-grams in the data model and not appearing in a new context, ignore the term;
- for a term comprising the one of the trending n-grams and not in the data model, check the term for frequency of use in the textual data and add the term in context to the training example if the frequency of use has reached a predetermined threshold;
- recompile the language model based on the term in context;
- and
- adopt the recompiled language model in an interactive virtual assistant.
8 more elements
Dependencies are inferred from the claim text.