GeLLM³O
GeLLM³O is an advanced large language model for multi-objective molecular optimization. It accepts molecular structures such as SMILES and descriptions of optimization objectives as input and outputs optimized candidate molecules.
Timeline (last 90 days)
GeLLM³O is an advanced LLM for multi-objective molecular optimization; it already possesses single-turn multi-objective optimization capability, but lacks an explicit training mechanism for utilizing historical feedback in multi-turn iteration.
The research team applied MFG to GeLLM³O, obtaining GeLLM³O-MFG.
Evaluated on ten multi-objective molecular optimization tasks, GeLLM³O-MFG achieved an average success rate of 81.50% under similarity constraints, significantly outperforming the base model GeLLM³O's 70.81%.
All Facts (3)
GeLLM³O is an advanced LLM for multi-objective molecular optimization; it already possesses single-turn multi-objective optimization capability, but lacks an explicit training mechanism for utilizing historical feedback in multi-turn iteration.
50%UnverifiedThe research team applied MFG to GeLLM³O, obtaining GeLLM³O-MFG.
50%UnverifiedEvaluated on ten multi-objective molecular optimization tasks, GeLLM³O-MFG achieved an average success rate of 81.50% under similarity constraints, significantly outperforming the base model GeLLM³O's 70.81%.
50%