{"id":"W4410193211","doi":"10.1038/s41467-025-59439-1","title":"Using GNN property predictors as molecule generators","year":2025,"lang":"en","type":"article","venue":"Nature Communications","topic":"Machine Learning in Materials Science","field":"Materials Science","cited_by":24,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"University of Toronto","keywords":"Property (philosophy); Molecule; Computer science; Chemistry","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000681374,0.0001503554,0.0001671931,0.0001336185,0.000722077,0.0002088335,0.002299873,0.0002103339,0.0002853711],"category_scores_gemma":[0.0008470001,0.0001085355,0.00004665452,0.0005941329,0.0003222675,0.0002040106,0.000922065,0.0005567437,0.0001428045],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001019112,"about_ca_system_score_gemma":0.0002826534,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002174,"about_ca_topic_score_gemma":0.00008366047,"domain_scores_codex":[0.9984685,0.000413119,0.0002813402,0.0003177475,0.0002609034,0.0002583634],"domain_scores_gemma":[0.9974466,0.0001271768,0.000120767,0.002051365,0.0001863648,0.00006770516],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.000007064117,0.00006884121,0.003567124,0.0000142348,0.000007296974,9.839596e-7,0.0001412898,0.0009031549,0.9695809,0.02193216,0.003634434,0.0001425321],"study_design_scores_gemma":[0.0005921668,0.00007772177,0.007333535,0.0002422677,0.0001147746,0.00003288386,0.0001803611,0.07032661,0.6714104,0.002900298,0.2460303,0.0007586739],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9429482,0.002532028,0.002949165,0.00543185,0.001796309,0.0005392129,0.00003200785,0.0004743505,0.04329684],"genre_scores_gemma":[0.9262604,0.0000456247,0.07133216,0.001206064,0.00004606484,0.00003562625,0.00001565866,0.00001438131,0.00104409],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2981705,"threshold_uncertainty_score":0.5553706,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02368592745329986,"score_gpt":0.3421802638476321,"score_spread":0.3184943363943322,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}