{"id":"W4415254285","doi":"10.48550/arxiv.2509.17398","title":"Optimizing Split Federated Learning with Unstable Client Participation","year":2025,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Beijing Jiaotong University; Tsinghua University; University of Hong Kong; Nanyang Technological University; Institute for Catastrophic Loss Reduction","keywords":"Upload; Federated learning; Computation; Convergence (economics); Rendering (computer graphics); Upper and lower bounds; Optimization problem; Edge device","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004855277,0.001040115,0.001449525,0.0003414858,0.0006168713,0.001330431,0.002304612,0.001792269,0.002123636],"category_scores_gemma":[0.01256061,0.0005234673,0.0005600087,0.0006198774,0.001691848,0.002769463,0.002280293,0.002217188,0.0004963938],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001596898,"about_ca_system_score_gemma":0.002390536,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003293632,"about_ca_topic_score_gemma":0.003366955,"domain_scores_codex":[0.9980522,0.0007773286,0.00007456453,0.0004337801,0.00028158,0.0003804759],"domain_scores_gemma":[0.9943699,0.003495001,0.0003724947,0.0008134565,0.0005561827,0.0003930034],"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":[0.000338389,0.0001897461,0.001448917,0.0000413796,0.00002767481,0.00006252377,0.00006179447,0.9581546,0.0007632498,0.008296493,0.001594856,0.02902042],"study_design_scores_gemma":[0.00001333021,0.00001873812,0.00005034241,0.000001797996,0.000002145679,0.000007138757,0.000008641836,0.9950551,0.0002291409,0.004549116,0.00006302597,0.00000137999],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1488036,0.0002754989,0.844793,0.001178533,0.00004930425,0.0001382445,0.0001459744,0.001297457,0.003318479],"genre_scores_gemma":[0.9453341,0.00005349216,0.05223186,0.0002209939,0.00002497765,0.0001126208,0.0001416424,0.00007175375,0.001808529],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004855277,"threshold_uncertainty_score":0.02567744,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07952927557461463,"score_gpt":0.222107378034756,"score_spread":0.1425781024601414,"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."}}