{"id":"W4392308684","doi":"10.1039/d4dd00024b","title":"Learning conditional policies for crystal design using offline reinforcement learning","year":2024,"lang":"en","type":"article","venue":"Digital Discovery","topic":"Crystallization and Solubility Studies","field":"Materials Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal; Polytechnique Montréal; Mila - Quebec Artificial Intelligence Institute","funders":"Institut de Valorisation des Données; Compute Canada; Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Reinforcement learning; Reinforcement; Computer science; Artificial intelligence; Psychology; Social psychology","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":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0002090854,0.000141095,0.0001580851,0.00007830507,0.0002991735,0.001311986,0.00007758155,0.0000351074,0.0001350184],"category_scores_gemma":[0.0003586822,0.0001265263,0.0001071305,0.0001189591,0.0001201201,0.001395036,0.00009277223,0.00008005746,0.00003309462],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009544106,"about_ca_system_score_gemma":0.00009583471,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000009795048,"about_ca_topic_score_gemma":0.000001569401,"domain_scores_codex":[0.9989518,0.00002671957,0.0002622507,0.0002615932,0.000237109,0.0002605313],"domain_scores_gemma":[0.999486,0.0002805228,0.00005106906,0.00007181555,0.00006732255,0.00004324389],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00009443452,0.00004939498,0.0004515566,0.000237993,0.00005720222,0.000007815506,0.001769212,0.8034688,0.1720097,0.02047698,0.001025848,0.0003510583],"study_design_scores_gemma":[0.001481686,0.001285576,0.001085165,0.0006413479,0.0001394847,0.00006216919,0.01091834,0.7935725,0.02846921,0.01263867,0.1481071,0.001598702],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3111356,0.0001619017,0.6862246,0.00009107301,0.0002392838,0.0002437621,0.00009392366,0.0001891638,0.001620688],"genre_scores_gemma":[0.9910733,0.000009894007,0.0003791114,0.00005845977,0.0001568689,0.00002438698,0.0002474964,0.00002177989,0.008028707],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6858455,"threshold_uncertainty_score":0.9997247,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04786053706346439,"score_gpt":0.3042265212792106,"score_spread":0.2563659842157462,"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."}}