{"id":"W3112351343","doi":"10.1088/2632-2153/ac09d6","title":"Deep molecular dreaming: inverse machine learning for de-novo molecular design and interpretability with surjective representations","year":2021,"lang":"en","type":"preprint","venue":"Machine Learning Science and Technology","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canadian Institute for Advanced Research; Vector Institute; University of Toronto","funders":"Office of Naval Research; Tata Steel","keywords":"Chemical space; Artificial intelligence; Computer science; Machine learning; Interpretability; Cheminformatics; Representation (politics); Property (philosophy); Generative model; Generative grammar; Chemistry; Drug discovery; Biology; Computational chemistry; Bioinformatics","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.003107285,0.0003801077,0.0004832316,0.0009451031,0.0006427887,0.0006190412,0.001155362,0.0002556595,0.00000307444],"category_scores_gemma":[0.004754788,0.0003681239,0.00006880197,0.001811812,0.001301673,0.0004361233,0.003098955,0.001554852,5.5647e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002032888,"about_ca_system_score_gemma":0.0009387878,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002220832,"about_ca_topic_score_gemma":0.00005572634,"domain_scores_codex":[0.9961541,0.0006794524,0.0003373469,0.00175154,0.0005097397,0.0005677721],"domain_scores_gemma":[0.9972778,0.0008146919,0.0003066241,0.0007162208,0.000732301,0.0001523599],"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":[0.00004487358,0.00008753499,0.02829031,0.0001347401,0.0001082243,0.0001770903,0.002642052,0.881771,0.01550766,0.01846255,9.769377e-7,0.05277294],"study_design_scores_gemma":[0.0003922801,0.0003110556,0.001143457,0.00008143596,0.0000446768,0.0002541164,0.0003062271,0.9606831,0.009533786,0.02682641,0.00004426641,0.0003791538],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3341245,0.0006699827,0.6630161,0.001342431,0.0000669818,0.000473777,0.000001443698,0.0002438562,0.00006089237],"genre_scores_gemma":[0.6293418,0.00002839587,0.37038,0.00008409587,0.000005691007,0.0001155674,0.00001125781,0.000018978,0.00001418322],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.2952172,"threshold_uncertainty_score":0.9998771,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01320056893517005,"score_gpt":0.2976142649902339,"score_spread":0.2844136960550638,"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."}}