{"id":"W4309208755","doi":"10.48550/arxiv.2211.08110","title":"HeatViT: Hardware-Efficient Adaptive Token Pruning for Vision Transformers","year":2022,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"CCD and CMOS Imaging Sensors","field":"Engineering","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; National Science Foundation","keywords":"Computer science; Security token; Computation; Quantization (signal processing); Computer hardware; Hardware acceleration; Edge device; Field-programmable gate array; Computer engineering; Latency (audio); Algorithm; Computer network","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0001670129,0.0003774662,0.0003664164,0.0002740909,0.0002386496,0.0000472076,0.0004326536,0.0001695344,0.0001522162],"category_scores_gemma":[0.000009509065,0.0004812683,0.0003543963,0.0003136577,0.00006701444,0.00009893328,0.0002288964,0.0006709592,0.0000172514],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005515753,"about_ca_system_score_gemma":0.00008210084,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006605274,"about_ca_topic_score_gemma":0.00001255622,"domain_scores_codex":[0.9985288,0.00004288314,0.0002010088,0.0006645722,0.0001023865,0.0004603281],"domain_scores_gemma":[0.9992431,0.00008885053,0.00006195431,0.0004006617,0.00006347137,0.0001419542],"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":[0.00008258189,0.00004278755,0.00007838442,0.0001648418,0.0001324685,0.00005428311,0.000616746,0.9947672,0.0001675978,0.001933522,0.000889681,0.001069917],"study_design_scores_gemma":[0.0006893414,0.00008689357,0.0001988637,0.0001133946,0.0001635399,0.000002742852,0.001034192,0.9831116,0.0003543881,0.0005870261,0.0130493,0.0006087164],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6281025,0.0002648187,0.3554824,0.00006602876,0.001650883,0.001140464,0.0003519717,0.0009520553,0.01198885],"genre_scores_gemma":[0.9976573,0.0001333608,0.0003888834,0.00003046615,0.00007778247,0.000007078565,0.00009764853,0.00008329321,0.001524181],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3695548,"threshold_uncertainty_score":0.9997639,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04724298686371685,"score_gpt":0.1854540897693705,"score_spread":0.1382111029056537,"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."}}