{"id":"W4389169645","doi":"10.1109/lgrs.2023.3338061","title":"Weakly Supervised Learning for Pixel-Level Sea Ice Concentration Extraction Using AI4Arctic Sea Ice Challenge Dataset","year":2023,"lang":"en","type":"article","venue":"IEEE Geoscience and Remote Sensing Letters","topic":"Arctic and Antarctic ice dynamics","field":"Earth and Planetary Sciences","cited_by":9,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada; Environment and Climate Change Canada","keywords":"Ground truth; Pixel; Synthetic aperture radar; Sea ice; Computer science; Polygon (computer graphics); Remote sensing; Artificial intelligence; Convolutional neural network; Image resolution; Geology; Climatology","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":[],"consensus_categories":[],"category_scores_codex":[0.0007485059,0.0002136099,0.0002010786,0.000119322,0.001164175,0.0001827577,0.0001456299,0.00009348939,0.00001402022],"category_scores_gemma":[0.0001167044,0.0001966775,0.00005645592,0.0004508543,0.0003021261,0.0006924842,0.00001737623,0.0002788778,0.00003943675],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002298671,"about_ca_system_score_gemma":0.00008277468,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006009853,"about_ca_topic_score_gemma":0.0008167884,"domain_scores_codex":[0.9980228,0.0001137984,0.000262563,0.0005555796,0.0003817154,0.0006635464],"domain_scores_gemma":[0.9990743,0.0003737085,0.0001402414,0.0001883914,0.00005915388,0.0001641749],"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":[0.0002604068,0.00003225121,0.01827701,0.000477179,0.00008264862,0.0002143833,0.004636582,0.05074499,0.1569101,0.00001199469,0.002373391,0.7659791],"study_design_scores_gemma":[0.0002998884,0.00007711949,0.01097707,0.00008287949,0.00004399152,0.000122127,0.0009303935,0.983269,0.0001713007,0.00006487547,0.003673876,0.000287497],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8802171,0.0000298589,0.1144699,0.003527177,0.00111846,0.0002680248,0.0002366823,0.00008320316,0.00004963025],"genre_scores_gemma":[0.9804647,0.0002641318,0.01559937,0.002054551,0.0003684697,8.047507e-8,0.001138803,0.00001206667,0.00009783955],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.932524,"threshold_uncertainty_score":0.908514,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04727062796431929,"score_gpt":0.2667448310820422,"score_spread":0.2194742031177229,"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."}}