{"id":"W4414271031","doi":"10.1109/prml66062.2025.11159676","title":"Molecular Structure Learning with Graph Transformers: A Graph Reduction Approach for Improved Efficiency and Accuracy","year":2025,"lang":"en","type":"article","venue":"","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"HORIZON EUROPE Health; Agencia Estatal de Investigación","keywords":"Overfitting; Pooling; Graph; Embedding; Molecular graph; Transformer; Training set; Graph reduction","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006171818,0.001008979,0.0006217091,0.001322588,0.0003477271,0.0007393319,0.001410913,0.00070899,0.004150499],"category_scores_gemma":[0.00283765,0.0003927336,0.0009685945,0.001158734,0.0006156233,0.002309068,0.001092324,0.001557475,0.001640189],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007872573,"about_ca_system_score_gemma":0.001210915,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003730439,"about_ca_topic_score_gemma":0.008484894,"domain_scores_codex":[0.9997039,0.0000867379,0.00001779867,0.00008542605,0.00007738775,0.00002881877],"domain_scores_gemma":[0.9991774,0.0003739038,0.00005527518,0.0002324548,0.0001240859,0.00003689232],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002240434,0.0002617261,0.001495849,0.0003035624,0.00008749118,0.0002021629,0.0001475011,0.28769,0.03798307,0.04062611,0.008966876,0.6220116],"study_design_scores_gemma":[0.00001651456,0.00006412112,0.0001291088,0.000008727331,0.00002185995,0.00005530254,0.00002412418,0.9597069,0.01175928,0.02591302,0.002291195,0.000009864232],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0238766,0.0001876495,0.9673141,0.0003262058,0.00004393714,0.0001046281,0.0002925588,0.005991518,0.001862839],"genre_scores_gemma":[0.3951789,0.0004122129,0.5957942,0.0003231251,0.0000484592,0.000188411,0.001792454,0.0009475088,0.005314731],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004150499,"threshold_uncertainty_score":0.01388484,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.007817234613003808,"score_gpt":0.2680013905701102,"score_spread":0.2601841559571064,"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."}}