{"id":"W4415108044","doi":"10.1175/aies-d-24-0107.1","title":"Adjustment of Decadal Ocean Carbon Sink Predictions Using Deep Learning","year":2025,"lang":"en","type":"article","venue":"Artificial Intelligence for the Earth Systems","topic":"Atmospheric and Environmental Gas Dynamics","field":"Environmental Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Environment and Climate Change Canada","funders":"","keywords":"Mean squared error; Artificial neural network; Anomaly (physics); Correlation coefficient; Carbon sink; Forecast skill; Climate model; Flexibility (engineering)","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003384856,0.0001421486,0.000173151,0.00001291005,0.0003180775,0.00002531259,0.0002439116,0.00007331832,0.00005716342],"category_scores_gemma":[0.00004349539,0.0001111468,0.00009346008,0.0002724539,0.0003045329,0.00007142674,0.0001170917,0.0001278274,0.00001358114],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001277807,"about_ca_system_score_gemma":0.00001186489,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001994225,"about_ca_topic_score_gemma":0.0001498868,"domain_scores_codex":[0.9987459,0.00006040361,0.0004466745,0.0002454554,0.0002281825,0.0002733798],"domain_scores_gemma":[0.9993777,0.0001718885,0.0001485416,0.0002425511,0.00001041199,0.00004894124],"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":[0.00002833684,0.00004888053,0.008703619,0.00001531887,0.00003467753,3.368261e-7,0.0003555319,0.9696487,0.001425173,0.001921222,0.000008036318,0.01781022],"study_design_scores_gemma":[0.00002752215,0.00009088374,0.001938073,0.00003530669,0.00006796422,0.000003029879,0.00330431,0.9905264,0.001794908,0.0006666286,0.001440344,0.0001046368],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4128466,0.0004222204,0.5840386,0.00005182245,0.0009184038,0.0006605457,0.000002955406,0.00003494765,0.001023867],"genre_scores_gemma":[0.9971314,0.00005796405,0.001649146,0.0000221403,0.00007826594,0.00002236151,0.000002209382,0.00001377662,0.001022706],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5842848,"threshold_uncertainty_score":0.4532435,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02235176106563529,"score_gpt":0.2546006528642195,"score_spread":0.2322488917985842,"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."}}