User Intent Identification via Neural Network Word Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional user intent identification systems in AI-based Human Machine Interfaces (HMIs) often fail to accurately capture the granular components of user intent, leading to incomplete understanding and incorrect extraction of user actions, as they rely on identifying key information within contiguous sequences of text, missing essential actions and additional user-provided context.
Innovation Solution
A method and system that processes user statements using Natural Language Understanding (NLU) to extract features such as Parts of Speech (POS) labels, dependency parse trees, and word embeddings, which are then used by a neural network model to predict classes for each word, enabling the identification of user intent as a set of inter-dependent components including actions and attributes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional systems capture user intent as a single or contiguous sequence of text, then the system structure remains simple, but the intent extraction becomes incomplete and inaccurate
Solution Approach 1:
The patent segments user intent into multiple granular components including action, centre of interest, and additional information, rather than treating it as a single contiguous text sequence. This segmentation allows each component to be extracted and analyzed separately, improving overall intent understanding accuracy while managing complexity through structured decomposition
Solution Approach 2:
The patent introduces a new dimensional approach by breaking down intent into vertical layers (action, centre of interest, additional information) rather than horizontal contiguous text sequences. This multi-dimensional classification framework enables more precise intent capture by considering multiple aspects simultaneously
2Reliability
If conventional systems focus on identifying key information in incoming queries, then processing speed is maintained, but critical action entities are missed leading to incorrect intent understanding
Solution Approach 1:
The patent applies preliminary action by first identifying and extracting the action component before determining the centre of interest and additional information. This sequential extraction approach ensures that critical action entities are captured first, preventing information loss and establishing a foundation for accurate overall intent understanding
Solution Approach 2:
The patent introduces an intermediary classification framework that acts as a mediator between raw user input and final intent understanding. This framework includes intermediate categories (action, centre of interest, additional information) that systematically bridge the gap between input text and accurate intent interpretation, preventing critical information loss
3Measurement precision
If conventional systems do not capture additional user-provided information, then processing complexity is reduced, but the understanding of true user intent becomes incomplete
Solution Approach 1:
The patent segments the intent structure into distinct components, with a specific category for additional information provided by users. This segmentation allows the system to systematically capture and process supplementary information without overwhelming complexity, as each segment has a defined role in the overall intent understanding
Solution Approach 2:
The patent applies local quality by treating different components of user input with different levels of analysis. Additional information provided by users is captured and processed with appropriate attention, allowing the system to focus computational resources on specific high-value information elements rather than uniformly processing all input
Data Source
AI summary
The present disclosure discloses method and a user intent identification system for identifying user intent from user statements. The user intent identification system receives input statement provided by a user from a Natural Language Understanding (NLU) engine. The input statement is processed to remove one or more irrelevant content. A plurality of features for each word in the processed input statement is extracted. The plurality of features comprises Parts of Speech (POS) label, dependency parse tree and word embeddings. The user intent determination system predicts class for each word in the processed input statement from a plurality of predefined classes using a neural network model. The neural network model predicts class for each word based on input vector generated for the each word based on the plurality of features. Thereafter, the user intent is identified based on class predicted for each word in processed input statement.