AI Star - Pioneering the AI Discovery of Materials and Pharmaceuticals

Patent & Certification

AI Star's Excellent Technical Capabilities

Swift Matrix

Deep Discovery

Patent & Cert

AI Star's
Excellence in Technology

Protein-Drug
Interaction Prediction
AI

Data Inteligence
Pipeline

Multi Scale
Phase Field Simulation

MPPS
(Multi-Physics PINN Software)

AI-based Drug-Target Screening

Use self-collected and purified drug big data and our own ABCnet-based deep learning model to analyze the relationship between the target protein amino acid sequence and the ligand SMILES with precision.

Through this, we provide candidates with the optimal binding force.

Improvement in Effect Prediction Accuracy

Improve the prediction accuracy of the target protein and ligand by 89% to 94% compared to the existing model and provide a self-developed deep learning-based prediction model.

Nano Scale Phase Field Simulation

The nano and micro particle growth simulation software analyzes the lithium-ion battery anode particles thermodynamically, predicts the crystallization pattern, particle size distribution, and porosity of the first and second particles.

Furthermore, it systematically manages data on lithium-ion distribution, insertion efficiency, insertion active area, and mechanical stress, and provides precise insights through advanced analysis.

MPPs
(Multi-Physics PINN Software)

The material physics prediction solution using PINN (Physics-Informed Neural Network) complements the limitations of physical-based simulations through rapid inference.

Through integration of various physical-based modules such as DFT and MD, it supports the development of models optimized for the prediction target phenomena.

Improvement in Capacity Prediction Accuracy

The prediction of the capacity of the material used in the secondary battery and the capacity prediction according to the material synthesis method is improved by 90% to 97% compared to the existing model and provided a self-developed deep learning-based prediction model.

Through years of experience, we have applied for 5 patents and published 3 related papers (bioXiv, JKICS, IJCAL).

The system and method for predicting the three-dimensional structure of a target protein, including a deep learning module, and the method for predicting the interaction between a drug and a target protein

Patent - 2021 - 0013839

The method for predicting the inhibition of the causal protein of a drug-induced nervous disease using a deep learning model

Patent - 2021 - 0013835

The method for deriving drug candidates using predicted active
walls

Patent - 2020 - 0117448