Context Representation Vector Generation for User Impression Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for evaluating the effect of information distribution, such as advertisements or news, on users' impressions of a subject fail to properly assess how the information changes the user's impression of the subject, relying mainly on metrics like browsing frequency rather than actual impression impact.
Innovation Solution
A provision device and method that generates and compares distributed representations of contexts before and after information distribution, using relative connections to quantify changes in user impressions, allowing for a more accurate evaluation of how distribution information affects user perceptions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional evaluation methods using browsing frequency are used, then the evaluation process is simple, but the evaluation precision of user impression change is insufficient
Solution Approach 1:
The patent transforms the evaluation from simple browsing frequency counts to distributed representation vectors that capture semantic relationships between contexts. By changing the parameter from scalar frequency to vector-based contextual representation, the system achieves higher measurement precision in evaluating user impression changes while managing complexity through automated vector operations.
Solution Approach 2:
The patent replaces manual or simple metric-based evaluation with an automated semantic analysis system using distributed representations. This substitution enables the system to automatically capture and measure changes in user impressions through vector comparisons, eliminating the need for complex manual evaluation processes while achieving superior precision.
2Measurement precision
If distributed representation is used to evaluate user impression changes, then the evaluation precision is improved, but the computational complexity increases
Solution Approach 1:
The patent pre-computes and stores distributed representation vectors for various contexts before they are needed for evaluation. By preparing these semantic representations in advance, the system reduces the computational power required during actual evaluation, as the heavy lifting of semantic analysis is performed beforehand rather than in real-time during impression change measurement.
3Loss of information
If only browsing frequency is measured, then the data collection is simple, but the information completeness about user impression is insufficient
Solution Approach 1:
The patent adds a new dimension to the evaluation by transforming scalar browsing frequency data into multi-dimensional distributed representation vectors. This dimensional transformation captures semantic relationships and contextual information that simple frequency counts miss, significantly reducing information loss about user impressions while the automated vector processing manages the added complexity.
Data Source
AI summary
According to one aspect of an embodiment a provision device includes a generation unit that generates a distributed representation of each context on the basis of a relative connection that multiple contexts have. The provision device includes a provision unit that provides information representing a change between a distributed representation of a given context that is generated before distribution information about the given context is distributed and a distributed representation of the given context that is generated after the distribution information is distributed.