Artificial intelligence apparatus and chemical material search method thereof

A non-transitory computer-readable storage medium having instructions stored thereon, when executed by a processor, cause the processor to execute a method of searching a chemical material in an artificial intelligence (AI) apparatus, the AI apparatus comprising:

• a database configured to store datasets of chemical materials;
• a quantum processor (QPU) configured to execute quantum annealing to obtain low energy states of a fingerprint;
• and
• the processor configured to search for the target material from the database, wherein the processor is configured to:
• predict a fingerprint relating to a target characteristic by inputting fingerprints of the datasets into a pre-trained neural network model;
• extract sample data using the quantum annealing executed by the QPU by:
• (i) optimizing the fingerprint relating to the target characteristic based on a cost function of a fingerprint prediction model expressed in a linear-quadratic form to obtain an optimized fingerprint that gives the lowest cost function;
• and (ii) calculating an optimized cost function based on the optimized fingerprint;
• evaluate a feature importance from the extracted sample data;
• select high-level features based on the feature importance;
• search for the target material by extracting molecular structures corresponding to the target material from the datasets in the database based on the selected high-level features;
• wherein the extracting of the molecular structures comprises:
• extract a preset number of molecular structures with small Hamming distances, including a molecular structure with the smallest Hamming distance;
• and
• extract a preset number of molecular structures with large Hamming distances, including a molecular structure have the largest Hamming distance;
• obtain a first common molecular structure from the extracted molecular structures with the smallest Hamming distances;
• obtain a second common molecular structure from the extracted molecular structures with the largest Hamming distances;
• and
• obtain specific key structure information of the target characteristic for searching for the target material based on the obtained first common molecular structure and the obtained second common molecular structure.

A method of searching a chemical material in an artificial intelligence (AI) apparatus, the AI apparatus comprising:

• a database configured to store datasets of chemical materials;
• a quantum processor (QPU) configured to execute quantum annealing to obtain low energy states of a fingerprint;
• and
• the method for searching the chemical material in the database, comprising:
• predicting a fingerprint relating to a target characteristic by inputting fingerprints of the datasets into a pre-trained neural network model;
• extracting sample data using quantum annealing executed by the QPU by:
• (i) optimizing the fingerprint relating to the target characteristic based on a cost function of a fingerprint prediction model expressed in a linear-quadratic form to obtain an optimized fingerprint that gives the lowest cost function;
• and
• (ii) calculating an optimized cost function based on the optimized fingerprint;
• evaluating a feature importance from the extracted sample data;
• selecting high-level features based on the feature importance;
• searching for the target material by extracting molecular structures corresponding to the target material from the datasets in the database based on the selected high-level features;
• wherein the extracting of the molecular structures comprises:
• extracting a preset number of molecular structures with small Hamming distances, including a molecular structure with the smallest Hamming distance;
• and
• extracting a preset number of molecular structures with large Hamming distances, including a molecular structure have the largest Hamming distance;
• obtaining a first common molecular structure from the extracted molecular structures with the smallest Hamming distances;
• obtaining a second common molecular structure from the extracted molecular structures with the largest Hamming distances;
• and
• obtaining specific key structure information of the target characteristic for searching for the target material based on the obtained first common molecular structure and the obtained second common molecular structure.