New FRAGMENTA model accelerates drug discovery with limited data

Researchers have developed FRAGMENTA, an end-to-end generative model designed for drug lead optimization, particularly effective in scenarios with limited training data. The model incorporates LVSEF, a fragment-based generator that optimizes fragmentation and generation simultaneously, and an agentic system that translates expert feedback into updated generative objectives. In testing across small-data datasets, LVSEF demonstrated superior performance compared to state-of-the-art methods in extremely limited data settings and matched them at larger scales, while also training significantly faster. Iterative optimization using FRAGMENTA showed improvements in discovery yield, with a real-world deployment identifying nearly twice as many molecules with favorable docking scores. AI

IMPACT Accelerates drug discovery pipelines, especially in data-scarce environments, by improving molecule generation and optimization.

RANK_REASON The cluster contains a research paper detailing a new AI model for drug discovery. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• Connected Papers
• DagsHub
• FRAGMENTA
• Gotit.pub
• Hugging Face
• Human-Agent FRAGMENTA
• Litmaps
• LVSEF
• ScienceCast
• Yuto Suzuki

• paper
• model release
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →