{"id":"W4381786635","doi":"10.1109/wf-iot54382.2022.10152039","title":"TinyHAR: Benchmarking Human Activity Recognition Systems in Resource Constrained Devices","year":2022,"lang":"en","type":"article","venue":"","topic":"Context-Aware Activity Recognition Systems","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Edge device; Benchmarking; Convolutional neural network; Inference; Artificial intelligence; Activity recognition; Deep learning; Benchmark (surveying); Machine learning; Quantization (signal processing); Context (archaeology); Computer vision","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001543943,0.0001913292,0.0003131215,0.0004038097,0.0004803194,0.0003074531,0.0006783472,0.00005874818,0.0003116184],"category_scores_gemma":[0.0000356557,0.0002153425,0.00008377465,0.0008235213,0.00003828527,0.0007852306,0.0005433813,0.0003936839,0.00005344355],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002953788,"about_ca_system_score_gemma":0.00009199849,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001438888,"about_ca_topic_score_gemma":0.0005749163,"domain_scores_codex":[0.9971402,0.0008779142,0.0004147162,0.0006413819,0.0005527242,0.0003730716],"domain_scores_gemma":[0.9986744,0.0004123972,0.0002810258,0.0004654583,0.00007436,0.00009237422],"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":[0.00006381042,0.0009420852,0.0128932,0.0002381128,0.0001182263,0.0002762064,0.003303643,0.0005290908,0.0228261,0.005847384,0.005605522,0.9473566],"study_design_scores_gemma":[0.01180329,0.002403414,0.08439821,0.001194285,0.00008972704,0.002932681,0.01656732,0.4902794,0.009165115,0.003799368,0.3710303,0.006336922],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8731123,0.0001276014,0.05991182,0.0009055074,0.001081544,0.001213812,0.0000472913,0.0008005511,0.06279961],"genre_scores_gemma":[0.9985281,6.620575e-7,0.0004124286,0.0002785295,0.0001059197,0.0002575768,0.0000263975,0.00001392705,0.0003765179],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9410197,"threshold_uncertainty_score":0.8781412,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04938194676671396,"score_gpt":0.2672890150702518,"score_spread":0.2179070683035378,"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."}}