{"id":"W4416429575","doi":"10.1109/iccad66269.2025.11241002","title":"Hummingbird: A Smaller and Faster Large Language Model Accelerator on Embedded FPGA","year":2025,"lang":"","type":"article","venue":"","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"Chinese Academy of Sciences","keywords":"Field-programmable gate array; Hummingbird; Limiting; Inference; Bandwidth (computing); Scheme (mathematics); Booster (rocketry); Unification","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0001774684,0.0004079627,0.0003517438,0.0002038558,0.0004588893,0.000397678,0.0009290163,0.0001702671,0.0001290755],"category_scores_gemma":[0.00002844413,0.0003789058,0.0000933386,0.0008668668,0.00009135086,0.0005156495,0.001080843,0.0004240869,0.0001695908],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005981955,"about_ca_system_score_gemma":0.000108469,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000003869303,"about_ca_topic_score_gemma":0.00003573534,"domain_scores_codex":[0.9972733,0.00006652551,0.0004624182,0.001199461,0.0002721446,0.0007261576],"domain_scores_gemma":[0.9980397,0.0001785637,0.0001217181,0.001351838,0.00009924476,0.0002089456],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001199501,0.0009926964,0.0001736456,0.0001406453,0.0001465261,0.00004065068,0.0103075,0.01740133,0.007578857,0.6859871,0.02304362,0.2540675],"study_design_scores_gemma":[0.0009527795,0.00007124474,0.0001507308,0.00008206497,0.00002293174,0.000003150791,0.000368678,0.9845534,0.005207576,0.003685885,0.004485352,0.0004161845],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04580256,0.0005811472,0.9233052,0.002828792,0.0002432406,0.0007384236,0.00001630516,0.0002555409,0.02622881],"genre_scores_gemma":[0.9091253,0.0001255211,0.04457697,0.01031585,0.00009919005,0.0001230459,0.000003672051,0.00002868032,0.03560175],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9671521,"threshold_uncertainty_score":0.9998663,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02279458213259505,"score_gpt":0.3028497259613687,"score_spread":0.2800551438287737,"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."}}