Interest Determination System Using Context Vector Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing interest determination systems struggle to accurately assess a user's interest in words that do not appear in their dialog logs, limiting their ability to understand user preferences beyond explicitly mentioned terms.
Innovation Solution
The system employs a neural network-based approach that learns from vector information representing user interests, incorporating context words before and after an interest estimation word to generate a taste model, which is then used to determine interest in both appearing and non-appearing words, utilizing multiple neural networks for optimization and context vector generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the system uses words that appear in user dialog to determine interest, then the determination is based on explicit user input, but the system cannot determine interest in words that do not appear in the dialog
Solution Approach 1:
The patent introduces context vectors as intermediary representations that bridge the gap between explicit dialog words and implicit user interests. By converting surrounding context words into vectors and combining them with the target word vector, the system creates a composite representation that serves as a mediator to infer interest in words not explicitly mentioned in the dialog.
Solution Approach 2:
The system performs preliminary vectorization of context words surrounding the target word before making the interest determination. By pre-processing and combining context information into a composite vector representation, the system prepares the necessary features in advance to accurately determine interest even for words not appearing in the dialog.
2Device complexity
If the system uses only explicit dialog words for interest determination, then the processing is simple, but the understanding of user preferences is limited
Solution Approach 1:
The patent transitions from analyzing only explicit dialog words to incorporating contextual information by converting words into vector representations. This dimensional transformation allows the system to capture semantic relationships and contextual meanings, thereby preserving more user preference information while maintaining manageable processing complexity through vector space operations.
3Measurement precision
If the system incorporates context information from surrounding words, then the interest determination becomes more accurate, but the processing complexity increases
Solution Approach 1:
The patent replaces traditional text processing and analysis methods with vector-based neural network processing. By substituting mechanical text manipulation with vector space operations and neural network computations, the system achieves higher accuracy in interest determination while managing processing complexity through efficient vector mathematics and learned representations.
Data Source
AI summary
An interest determination system includes a data acquisition unit configured to acquire a first data group indicating a feature of each word contained in user's dialog information; a first learning unit configured to learn by using a first input and output values, the first input value being set to a data group that associates the first data group with a second data group indicating a feature of a predetermined word, the first output value being set to information indicating whether there is a user's interest; and a determination unit configured to, when a data group that associates a third data group indicating a feature of a word to be determined with the first data group has been input to the first learning unit, determine whether there is a user's interest in the word to be determined, based on an output value output from the first learning unit.