Intent Extraction Using Probabilistic and Deterministic Pattern Matching

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Machine learning systems face ambiguity and conflict in determining user intent due to syntactic variations in utterances, leading to incorrect intent outputs and the need for human intervention, additional training data, and resource-intensive update-tune-validate loops.

Innovation Solution

A system that combines probabilistic and deterministic machine learning processes, using probabilistic logic to determine candidate intents and deterministic logic with regular expression patterns to resolve ambiguities, allowing for semi-automated creation of regular expressions to encode utterance meaning directly within syntax, and providing feedback mechanisms to improve accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If probabilistic machine learning is used to determine user intent, then the system can handle syntactic variations, but ambiguity and conflict in intent determination increase

Engineering Contradiction:

Improvehandling syntactic variationsVSAvoidintent determination accuracy

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The system segments the intent determination process into two distinct modules: a probabilistic machine learning module for handling syntactic variations and a deterministic pattern matching module for resolving ambiguities. This segmentation allows each module to specialize in specific tasks, improving overall reliability while maintaining adaptability to different utterance forms.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

A threshold-based selection mechanism acts as an intermediary between the probabilistic and deterministic modules. When the probabilistic module identifies multiple candidate intents above the threshold, the system transfers control to the deterministic module for final resolution. This intermediary mechanism smoothly transitions between probabilistic and deterministic approaches, resolving the contradiction between adaptability and reliability.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If deterministic pattern matching is used to resolve intent ambiguity, then intent determination accuracy improves, but the system complexity increases

Engineering Contradiction:

Improveintent determination accuracyVSAvoidsystem architecture complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system merges probabilistic machine learning and deterministic pattern matching into a unified intent determination architecture. Both modules process the same input utterances and contribute to the final intent selection through a coordinated threshold-based mechanism. This merging approach improves reliability through complementary strengths while avoiding the complexity of implementing separate independent systems.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The deterministic pattern matching module is activated only partially - specifically when the probabilistic module identifies multiple candidate intents above the threshold. This partial activation strategy ensures that deterministic precision is applied only where needed, reducing overall system complexity while maintaining high reliability for ambiguous cases.

Inventive Principle:

Principle #16Partial or excessive action

3Reliability

If human intervention is used to resolve intent ambiguity, then intent determination accuracy improves, but the time and resources required increase

Engineering Contradiction:

Improveintent determination accuracyVSAvoidtime for human intervention

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system performs self-service by automatically resolving intent ambiguities through deterministic pattern matching without requiring human intervention. The deterministic module processes and resolves ambiguous cases autonomously, eliminating the time and resources that would otherwise be needed for human review and decision-making while maintaining high accuracy.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system incorporates feedback mechanisms where the deterministic module's resolutions can be used to refine future probabilistic predictions. This feedback loop allows the system to learn from and improve upon previous ambiguous cases, reducing the need for human intervention over time while maintaining high reliability in intent determination.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS11164564B2Augmented intent and entity extraction using pattern recognition interstitial regular expressions

Publication Date: 2021.11.02 BANK OF AMERICA CORP

AI summary

According to certain embodiments, a system comprises interface circuitry and processing circuitry. The processing circuitry receives an input via the interface circuitry. The input is based on an utterance of a user, and the processing circuitry uses a probabilistic engine to determine one or more candidate intents associated with the utterance. The processing circuitry determines a number of the one or more candidate intents that exceed a threshold. If the number of candidate intents that exceed the threshold does not equal one, the processing circuitry uses a deterministic engine to compare the input to a set of regular expression patterns. If the input matches one of the regular expression patterns, the processing circuitry uses the matching regular expression pattern to determine the intent of the utterance. The interface circuitry communicates the intent of the utterance as an output.