Idea Vector Search Balancing Candidate Quality and Novelty
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current artificial intelligence methods for generating candidate ideas rely heavily on initial conditions, leading to a lack of diversity and efficiency in idea generation, as they often suggest similar ideas based on past trials, without considering novelty or quality effectively.
Innovation Solution
A system that utilizes a multivariate space to generate candidate ideas by converting input idea vectors into contracted vectors, maintaining similarity between input and output vectors, and incorporating predetermined index values to ensure both novelty and quality are considered in the generation process, using a variable auto-encoder (VAE) to optimize the search for candidate ideas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an interpolation search method using artificial intelligence technology is used to generate candidate ideas, then the quality of candidate ideas based on known information is improved, but the diversity and novelty of candidate ideas deteriorate
Solution Approach 1:
The patent merges two search strategies: exploitation (interpolation search based on known information) and exploration (random search for novelty). The system combines results from both approaches to generate candidate ideas that balance quality and diversity, resolving the contradiction between using established methods and discovering new possibilities
Solution Approach 2:
The patent dynamically adjusts the balance between exploitation and exploration based on the search state. When the search space is well-explored, it emphasizes exploitation for quality; when diversity is needed, it increases exploration through random search components, making the system adaptable to different stages of the idea generation process
2Adaptability or versatility
If simple random search is used to generate candidate ideas, then the novelty of candidate ideas is improved, but the search efficiency deteriorates
Solution Approach 1:
The patent combines random search (for novelty) with interpolation search (for efficiency) into a hybrid approach. The system uses random search to generate diverse candidate ideas while employing interpolation search to efficiently evaluate and rank them based on known information, thus maintaining both novelty and search efficiency
Solution Approach 2:
The patent introduces an intermediary evaluation mechanism that uses known information to assess randomly generated ideas. This intermediary layer filters and ranks random candidates based on their potential quality, preventing purely random search while maintaining novelty, and efficiently identifying promising candidates from the random pool
Data Source
AI summary
A system converts a first input idea vector representing an idea into a first contracted vector. The system generates one or more second contracted vectors in a multivariate space to which the first contracted vector belongs, based on a value of a first predetermined component of the first contracted vector in the multivariate space. The system generates, respectively from the one or more second contracted vectors, one or more first output idea vectors representing a candidate idea to be proposed to a user. The multivariate space is configured to maintain a similarity between an input idea vector for generating a contracted vector and an output idea vector generated from the contracted vector and a similarity between the first predetermined component of the contracted vector and a first predetermined index value.