AI Chemical Material Search Using Quantum Annealing Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The development of new chemical materials with desired characteristics is time-consuming due to the vast chemical space and the need to examine numerous microscopic factors, necessitating more efficient search methods.
Innovation Solution
An artificial intelligence apparatus using quantum annealing and a pre-trained neural network model to predict fingerprints, extract sample data, evaluate feature importance, and search for target materials based on high-level features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used to examine vast combinations within chemical space, then comprehensive material evaluation is achieved, but development time becomes excessively long
Solution Approach 1:
The patent extracts and evaluates only the most critical features and materials from the vast chemical space using machine learning models. By identifying key descriptors that most strongly correlate with target properties, the system extracts a minimal subset of materials warranting detailed evaluation, thereby reducing development time while maintaining evaluation comprehensiveness for the most promising candidates.
Solution Approach 2:
The patent performs preliminary screening and feature importance analysis before detailed material evaluation. Machine learning models are trained beforehand to identify critical descriptors, and this preliminary action guides subsequent focused experimentation on pre-selected candidate materials, avoiding exhaustive examination of all possible combinations.
2Productivity
If algorithms are used to interpret molecules and search for target materials, then search efficiency is improved, but search time remains considerable
Solution Approach 1:
The patent replaces traditional mechanical search methods with quantum annealing computation. By formulating the material search problem as a quadratic unconstrained binary optimization (QUBO) problem and solving it using quantum annealing, the system achieves exponential speedup in exploring chemical space compared to classical algorithms, significantly reducing search time while improving efficiency.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach allows for quick and efficient identification of chemical materials meeting desired characteristics by reducing the chemical space and optimizing search time.
Implementation Method 1
extract sample data by optimizing the fingerprint relating to the target characteristic
Data Source
AI summary
The present invention relates to an artificial intelligence apparatus and a chemical material search method thereof that are capable of efficiently searching, by means of an annealing-based quantum computing device, for a chemical material satisfying desired characteristics, wherein the apparatus comprises a database that stores datasets of a chemical material, and a processor that searches for a target material from the database, and the processor may predict fingerprints relating to target characteristics by inputting fingerprints of the datasets into a pre-trained neural network model, extract sample data by optimizing the fingerprints relating to target characteristics, evaluate feature importance from the extracted sample data, select high-level features on the basis of the feature importance, and search for a target material on the basis of the selected high-level features.