{"id":"W2995426144","doi":"10.48550/arxiv.2001.08248","title":"How Much Position Information Do Convolutional Neural Networks Encode?","year":2020,"lang":"en","type":"article","venue":"arXiv (Cornell University)","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":94,"is_retracted":false,"has_abstract":true,"ca_institutions":"Toronto Metropolitan University","funders":"","keywords":"ENCODE; Convolutional neural network; Computer science; Artificial intelligence; Set (abstract data type); Position (finance); Contrast (vision); Artificial neural network; Filter (signal processing); Deep learning; Pattern recognition (psychology); 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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000830633,0.0008593489,0.0006716349,0.0007890321,0.0004285419,0.002282587,0.0009625757,0.001170435,0.003130564],"category_scores_gemma":[0.0110417,0.0007350799,0.0004318517,0.001445201,0.001691936,0.009848103,0.0008514001,0.001468739,0.001854095],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008652362,"about_ca_system_score_gemma":0.0006911547,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003775492,"about_ca_topic_score_gemma":0.004149405,"domain_scores_codex":[0.9994142,0.0001111837,0.00003306289,0.0002164401,0.0001407476,0.00008432894],"domain_scores_gemma":[0.9976782,0.0008479826,0.0004126503,0.000621666,0.0003559444,0.00008352774],"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":[0.0005929575,0.0001290766,0.02068601,0.0007559162,0.0003249309,0.0002130901,0.0006131953,0.09729067,0.02539694,0.1826494,0.01436919,0.6569786],"study_design_scores_gemma":[0.00005936533,0.0002084807,0.01391709,0.0004634767,0.0002827644,0.0004252272,0.0005261873,0.2986225,0.020382,0.6373092,0.02768098,0.0001227764],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2426256,0.009493303,0.6928984,0.01765445,0.0009327154,0.00004832114,0.004033958,0.001824434,0.03048898],"genre_scores_gemma":[0.9468694,0.004811174,0.04107147,0.0008780131,0.0002530727,0.00004255519,0.001483034,0.0002311532,0.00436021],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003775492,"threshold_uncertainty_score":0.01047277,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03836477917555642,"score_gpt":0.1652807220127223,"score_spread":0.1269159428371658,"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."}}