{"id":"W4321194910","doi":"10.1109/lra.2023.3245405","title":"Towards More Efficient EfficientDets and Real-Time Marine Debris Detection","year":2023,"lang":"en","type":"article","venue":"IEEE Robotics and Automation Letters","topic":"Microplastics and Plastic Pollution","field":"Environmental Science","cited_by":54,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"National Science and Technology Council; Queen's University; Natural Environment Research Council; Queen's University Belfast; Sight Research UK","keywords":"Underwater; Computer science; Object detection; Environmental science; Debris; Detector; Latency (audio); Real-time computing; Remote sensing; Artificial intelligence; Computer vision; Marine engineering; Geology; Pattern recognition (psychology); Engineering; Oceanography; Telecommunications","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001747836,0.0001162075,0.0001007801,0.00007872042,0.0001796186,0.00005888725,0.00004617024,0.00004864619,0.00007955015],"category_scores_gemma":[0.00002557658,0.000108504,0.00002114445,0.0002513573,0.0001019378,0.00004906951,0.00008223297,0.00005641456,0.000236441],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006001926,"about_ca_system_score_gemma":0.000003774847,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000178877,"about_ca_topic_score_gemma":0.00001334057,"domain_scores_codex":[0.999169,0.00002422792,0.0001598299,0.0002309276,0.0002049873,0.0002110528],"domain_scores_gemma":[0.9996973,0.00005393926,0.00006324616,0.00009413232,0.000005473539,0.00008585357],"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":[0.000007622767,0.00002026244,0.001068147,0.00001831624,0.000008550016,0.000008855931,0.0002680472,0.5176042,0.4602655,0.0000324602,0.001440422,0.01925761],"study_design_scores_gemma":[0.0002031177,0.00002501524,0.1849874,0.00001047079,0.00001673768,0.00001424178,0.00001036607,0.8110449,0.003348516,0.00002320446,0.0001882115,0.0001278092],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9684142,0.000002276635,0.02976071,0.001088671,0.0002492639,0.0001222602,0.000009017103,0.0001328472,0.0002206979],"genre_scores_gemma":[0.9982162,0.00003544593,0.001397732,0.0001879646,0.00003672853,0.000004297749,0.00001506813,0.00001171474,0.00009481677],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.456917,"threshold_uncertainty_score":0.4424664,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.005751159098058478,"score_gpt":0.2010989907747207,"score_spread":0.1953478316766622,"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."}}