{"id":"W4283719127","doi":"10.1101/2022.06.26.497684","title":"Improving the workflow to crack Small, Unbalanced, Noisy, but Genuine (SUNG) datasets in bioacoustics: the case of bonobo calls","year":2022,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Animal Vocal Communication and Behavior","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université du Québec à Chicoutimi","funders":"LabEx ASLAN; Université de Lyon; Social Sciences and Humanities Research Council of Canada; Institut Universitaire de France; Agence Nationale de la Recherche","keywords":"Bonobo; Computer science; Workflow; Artificial intelligence; Robustness (evolution); Cluster analysis; Machine learning; Feature vector; Support vector machine; Feature (linguistics); Pattern recognition (psychology); Biology","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.000906104,0.0003988774,0.0003577266,0.00009841016,0.0002451829,0.00009769831,0.001599062,0.000321002,0.00004823103],"category_scores_gemma":[0.0004305669,0.0003082896,0.0001388883,0.0003880576,0.0001998986,0.000006241893,0.002998543,0.000848979,0.000009849738],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001041015,"about_ca_system_score_gemma":0.0003953143,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007718696,"about_ca_topic_score_gemma":0.0002561133,"domain_scores_codex":[0.9977626,0.0003061868,0.0005802967,0.0007396308,0.0001891011,0.0004222216],"domain_scores_gemma":[0.9963993,0.00008069413,0.000382995,0.00276637,0.000220598,0.0001500511],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00009698827,0.0001819613,0.001701611,0.000089359,0.00006360812,0.0001410138,0.00001299145,0.0004007266,0.9966254,0.00006749488,0.0005590654,0.00005974971],"study_design_scores_gemma":[0.001789385,0.0007377894,0.07264683,0.0003351554,0.0006112441,0.000002791028,0.0002568456,0.001945293,0.8658808,0.000004067396,0.05333782,0.002451972],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9933807,0.001694118,0.001114936,0.0004839629,0.0003228417,0.0009721944,0.001994244,0.00003282896,0.000004147632],"genre_scores_gemma":[0.9953585,0.0003174742,0.002955264,0.0007217595,0.0002089716,0.0003367214,0.00001585507,0.00007365613,0.00001174225],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1307446,"threshold_uncertainty_score":0.9999369,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01996363930098752,"score_gpt":0.2550388992871567,"score_spread":0.2350752599861692,"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."}}