{"id":"W4413948581","doi":"10.26434/chemrxiv-2025-wd80w","title":"LightChem: A Lightweight Domain-Specific Language Model for Molecular Property and Reaction Prediction in Chemistry","year":2025,"lang":"en","type":"article","venue":"ChemRxiv","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Ministry of Education and Child Care","funders":"National Science and Technology Major Project; Chinese Academy of Sciences; National Natural Science Foundation of China","keywords":"Property (philosophy); Domain (mathematical analysis); Chemistry; Computer science; Mathematics; Philosophy; Epistemology","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":[0.00119987,0.001278218,0.0008336393,0.001009732,0.0005912962,0.002556233,0.003686185,0.001729335,0.009546162],"category_scores_gemma":[0.003167668,0.0009309807,0.002864296,0.0006794261,0.0006976736,0.002711578,0.001838579,0.002317522,0.005144394],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001450706,"about_ca_system_score_gemma":0.003000111,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007562514,"about_ca_topic_score_gemma":0.008897295,"domain_scores_codex":[0.9994547,0.0001065837,0.0000678104,0.0001421867,0.0001745303,0.00005422396],"domain_scores_gemma":[0.9990481,0.000494174,0.00007779257,0.0002093288,0.000117082,0.00005360446],"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":[0.000726881,0.0004224021,0.004551881,0.001996414,0.0003530924,0.0007796004,0.0005915429,0.5704898,0.01731405,0.200266,0.07733493,0.1251734],"study_design_scores_gemma":[0.0001313541,0.00005393977,0.0002068622,0.00008344123,0.00005876302,0.0001138809,0.00002936696,0.8566777,0.00728027,0.04332156,0.09199309,0.00004972294],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.007005733,0.0003687603,0.8896329,0.000531154,0.000126131,0.0003503249,0.01282755,0.08213352,0.007023853],"genre_scores_gemma":[0.148162,0.001186906,0.7888789,0.001057888,0.0001147968,0.002193819,0.0366463,0.01113611,0.01062338],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009546162,"threshold_uncertainty_score":0.0319351,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01314849883223688,"score_gpt":0.2643622996448124,"score_spread":0.2512138008125755,"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."}}