{"id":"W4403320045","doi":"10.1101/2024.10.07.617002","title":"Improving the reliability of molecular string representations for generative chemistry","year":2024,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Scientific Computing and Data Management","field":"Decision Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"Grand Équipement National De Calcul Intensif; Agence Nationale de la Recherche","keywords":"Generative grammar; String (physics); Reliability (semiconductor); Chemistry; Computer science; Computational biology; Computational chemistry; Theoretical physics; Artificial intelligence; Biology; Physics; Thermodynamics","routes":{"ca_aff":true,"ca_fund":false,"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":["metaresearch","metaepi_narrow","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.007442995,0.0003364863,0.0004599863,0.0002065659,0.0002802497,0.001087547,0.002074582,0.0001942066,0.00004075482],"category_scores_gemma":[0.00836946,0.0002486525,0.0003468399,0.001135318,0.0002490756,0.0001085689,0.003474321,0.0005391815,0.00002789487],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000146299,"about_ca_system_score_gemma":0.0005586833,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000831738,"about_ca_topic_score_gemma":0.000001330308,"domain_scores_codex":[0.9951002,0.0002012659,0.001133826,0.001953261,0.001230567,0.0003808522],"domain_scores_gemma":[0.9925526,0.0009686613,0.000758733,0.004243067,0.001347271,0.0001296633],"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":[0.00002270135,0.0001102037,0.0006020141,0.000648807,0.0001559004,0.00001475366,0.00006684436,0.01307127,0.9742573,0.003631052,0.007335495,0.00008359396],"study_design_scores_gemma":[0.0002837609,0.0000227775,0.003228795,0.0002488342,0.0002307515,9.992582e-9,0.00008868355,0.07788254,0.9116905,0.001119988,0.004639041,0.0005643638],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8686404,0.0006985507,0.1247363,0.001027146,0.00260106,0.001217372,0.0008416338,0.0001908307,0.00004669372],"genre_scores_gemma":[0.9840654,0.000007286307,0.01529191,0.00005714703,0.0002675572,0.0002040749,7.018245e-7,0.00004110838,0.00006483408],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.115425,"threshold_uncertainty_score":0.9999965,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04636439484916314,"score_gpt":0.315680954438788,"score_spread":0.2693165595896249,"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."}}