Back to results
US 12,450,505 B2Granted
Utilizing large language model responses to train an inference pattern engine
AssigneeTiny Fish Inc.
InventorsShuhao Zhang, Qi Zhai, Pavlo Dudka, Mark Yi Chien Peng, FNU Taha Junaid
Life of this patent
Estimating the term…
Abstract
A prompt is generated based on an input and processed webpage or application content. The prompt is provided to a large language model. Based on a response received from the large language model that maps one or more variables associated with the input to one or more elements associated with the processed content associated with the tree data structure received from the large language, a pattern mapping the one or more variables associated with the input to the one or more elements associated with the processed content associated with the tree data structure is derived.
Claims20
A system, comprising:
- a processor configured to:
- generate a plurality of prompts based on a plurality of inputs and processed content associated with a tree data structure;
- provide the plurality of prompts to a large language model;
- based on a response received from the large language model that maps one or more variables associated with the plurality of inputs to one or more elements associated with the processed content associated with the tree data structure received from the large language model, derive corresponding patterns mapping one or more variables associated with the plurality of inputs to the one or more elements associated with the processed content associated with the tree data structure;
- store the derived pattern mappings and a corresponding correctness associated with the derived pattern mappings based on the large language model response;
- receive a subsequent input and subsequent processed content associated with the tree data structure;
- determine that a correctness associated with a stored derived pattern mapping from the derived pattern mappings is greater than a confidence threshold;
- and
- utilize the stored derived pattern mapping to generate a second response for the subsequent input;
- and
- a memory coupled to the processor and configured to provide the processor with one or more instructions.
8 more elements
A method, comprising:
- generating a plurality of prompts based on a plurality of inputs and processed content associated with a tree data structure;
- providing the plurality of prompts to a large language model;
- based on a response received from the large language model that maps one or more variables associated with the plurality of inputs to one or more elements associated with the processed content associated with the tree data structure received from the large language model, deriving corresponding patterns mapping one or more variables associated with the plurality of inputs to the one or more elements associated with the processed content associated with the tree data structure;
- storing the derived pattern mappings and a corresponding correctness associated with the derived pattern mappings based on the large language model response;
- receiving a subsequent input and subsequent processed content associated with the tree data structure;
- determining that a correctness associated with a stored derived pattern mapping from the derived pattern mappings is greater than a confidence threshold;
- and
- utilizing the stored derived pattern mapping to generate a second response for the subsequent input.
5 more elements
A computer program product embodied in a non-transitory computer readable medium and comprising computer instructions for:
- generating a plurality of prompts based on a plurality of inputs and processed content associated with a tree data structure;
- providing the plurality of prompts to a large language model;
- based on a response received from the large language model that maps one or more variables associated with the plurality of inputs to one or more elements associated with the processed content associated with the tree data structure received from the large language model, deriving corresponding patterns mapping one or more variables associated with the plurality of inputs to the one or more elements associated with the processed content associated with the tree data structure;
- storing the derived pattern mappings and a corresponding correctness associated with the derived pattern mappings based on the large language model response;
- receiving a subsequent input and subsequent processed content associated with the tree data structure;
- determining that a correctness associated with a stored derived pattern mapping from the derived pattern mappings is greater than a confidence threshold;
- and
- utilizing the stored derived pattern mapping to generate a second response for the subsequent input.
5 more elements
Dependencies are inferred from the claim text.