Voice Command Item Identification Using Concept Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Ambiguous voice commands when ordering items online using voice assistants can lead to incorrect additions or removals from a cart due to similar titles of items, causing user intent misinterpretation.
Innovation Solution
A voice command system that generates a concept vector and feature vector for each item in the cart, calculates a candidate score based on these vectors, and identifies the intended item by determining the highest scoring item, thereby reducing ambiguity and improving user intent recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If voice commands are used to order items online, then ease of operation is improved, but measurement precision deteriorates due to ambiguous item identification
Solution Approach 1:
The system transforms the item identification problem from simple title matching to a multi-parameter comparison task. It generates concept vectors and feature vectors that capture multiple attributes of items (category, brand, attributes, etc.), then compares these vectors against the voice command intent to identify the correct item. This parameter expansion resolves the ambiguity that plagues simple title-based matching.
Solution Approach 2:
The system introduces vector representations as an intermediary between the voice command and the item catalog. Instead of directly matching voice commands to item titles, the system converts both into comparable vector forms (concept vectors for commands, feature vectors for items) and uses similarity scoring as a mediator to determine the best match. This intermediary transformation enables precise identification even when titles are similar.
2Device complexity
If simple title matching is used for voice commands, then device complexity is reduced, but reliability deteriorates due to ambiguous item identification
Solution Approach 1:
The system segments the item identification process into distinct computational stages: extracting features from item data, generating concept vectors from voice commands, computing similarity scores between vectors and items, and selecting the highest-scoring match. This segmentation allows each stage to be optimized independently while maintaining overall reliability, despite the increased complexity of using vector comparisons rather than simple string matching.
3Measurement precision
If vector-based item identification is implemented, then measurement precision is improved, but use of energy increases due to computational requirements
Solution Approach 1:
The system computes similarity scores for all items in the cart against the voice command concept vector, which may be more computation than strictly necessary if only the top few matches are needed. However, this exhaustive approach ensures that the highest-scoring item is correctly identified even in edge cases with similar items, prioritizing reliability over minimal computational effort.
Data Source
AI summary
A system including one or more processors and one or more non-transitory computer-readable media storing computing instructions configured to run on the one or more processors and perform receiving a voice command from a user to perform a virtual action intended to apply to one item of two or more items in a cart of the user; generating a concept vector representing a concept in the voice command; transforming the respective item attributes for each of the two or more items into a respective feature vector; generating a respective candidate score for the each of the two or more items; identifying the one item to which the voice command is intended to apply; and executing an action with respect to the one item based on the voice command. Other embodiments are disclosed.