Neural Network for Chemical Compound Structural Feature Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current neural networks struggle to effectively capture structural features of chemical compounds due to their variable length representations, limiting their ability to accurately classify and analyze chemical structures.
Innovation Solution
A computer-implemented method is introduced to train a neural network by reading and converting chemical compound expressions into structured training data, enumerating position and type combinations, and generating values indicating the presence of structural elements, allowing the network to capture specific structural features of chemical compounds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If variable length expressions are used to represent chemical compounds, then the neural network can accommodate diverse chemical structures, but the network's ability to accurately capture structural features deteriorates
Solution Approach 1:
The patent segments the chemical compound representation into fixed-length positional bins (e.g., 128 bins) where each bin captures the presence and type of structural elements at specific positions. This segmentation transforms variable-length molecular expressions into fixed-length vectors, enabling the neural network to process diverse chemical structures uniformly while maintaining structural feature accuracy through position-aware encoding.
2Productivity
If conventional neural networks are used for classification, then processing speed is maintained, but the precision in identifying specific structural features deteriorates
Solution Approach 1:
The patent introduces a positional dimension to the neural network input by organizing structural element frequencies into position-based bins. This dimensional transformation allows the network to simultaneously process structural composition and spatial arrangement information, improving structural feature identification precision while maintaining processing efficiency through vectorized operations.
Data Source
AI summary
A computer implemented method for training a neural network to capture a structural feature specific to a set of chemical compounds is disclosed. In the method, the computer system reads an expression describing a structure of the chemical compound for each chemical compound in the set and enumerates one or more combinations of a position and a type of a structural element appearing in the expression for each chemical compound in the set. The computer system also generates training data based on the one or more enumerated combinations for each chemical compound in the set. The training data includes one or more values with a length, each of which indicates whether or not a corresponding type of the structural element appears at a corresponding position for each combination. Furthermore, the computer system trains the neural network based on the training data for the set of the chemical compounds.