BLOX

BLOX performs objective-free exploration of chemical property space to identify compounds with novel materials and optical properties.

Key Features:

• Objective-Free Exploration: Operates without predefined objectives or boundaries in property space to enable open-ended search.
• Kernel-Based Stein Discrepancy: Uses a criterion based on kernel-based Stein discrepancy to assess discrepancies in property space for candidate selection.
• Minimization of DFT Calculations: Minimizes the number of density functional theory (DFT) calculations required to reduce computational cost.
• Intensity–Wavelength Property Space: Targets complex property spaces such as intensity–wavelength for optical-property exploration.
• Experimental Validation via Absorption Spectroscopy: Supports identification of light-absorbing molecules that were validated experimentally by absorption spectroscopy.

Scientific Applications:

• Identification of Light-Absorbing Molecules: Identified light-absorbing molecules from a drug database and validated eight compounds with outstanding optical properties experimentally.
• Chemical Repurposing: Facilitates repurposing of existing drug compounds based on discovered optical properties.
• Discovery of Out-of-Trend Compounds: Enables finding compounds that diverge from trends in complex property spaces such as intensity–wavelength.
• Materials and Renewable Energy Discovery: Applicable to materials chemistry challenges including renewable energy materials requiring novel optical or functional properties.

Methodology:

Objective-free exploration guided by a kernel-based Stein discrepancy criterion and strategies to minimize density functional theory (DFT) calculations.