{"id":"W4399332322","doi":"10.1016/j.neunet.2024.106420","title":"Capacity bounds for hyperbolic neural network representations of latent tree structures","year":2024,"lang":"en","type":"article","venue":"Neural Networks","topic":"Topological and Geometric Data Analysis","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"McMaster University; Vector Institute","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Embedding; Mathematics; Representation (politics); Tree (set theory); Distortion (music); Euclidean space; Curvature; Dimension (graph theory); Artificial neural network; Upper and lower bounds; Activation function; Hyperbolic tree; Function (biology); Graph; Perceptron; Euclidean geometry; Discrete mathematics; Algorithm; Combinatorics; Computer science; Hyperbolic geometry; Pure mathematics; Artificial intelligence; Differential geometry; Mathematical analysis; Geometry","routes":{"ca_aff":true,"ca_fund":true,"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":[0.005599329,0.001354331,0.001843746,0.002267093,0.001010695,0.003671427,0.004023823,0.002447428,0.01011592],"category_scores_gemma":[0.03777512,0.001087782,0.00107115,0.002785628,0.003808973,0.01162486,0.006160113,0.006066886,0.001071555],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00366038,"about_ca_system_score_gemma":0.001761028,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004258075,"about_ca_topic_score_gemma":0.004722944,"domain_scores_codex":[0.9978966,0.0008125442,0.0001256817,0.0003839091,0.0004569977,0.0003242333],"domain_scores_gemma":[0.9587469,0.0337956,0.001578857,0.002983821,0.001778313,0.001116629],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001879743,0.00008531392,0.0008500463,0.0003114666,0.00007481716,0.00007616507,0.0002820352,0.2980275,0.001276395,0.6671858,0.005445531,0.02619693],"study_design_scores_gemma":[0.000007475316,0.00001087371,0.0001292928,0.00004515572,0.0000104737,0.00001900354,0.0000265308,0.615615,0.0003179204,0.3833891,0.0004133387,0.00001587342],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.07570048,0.002182994,0.901233,0.003534198,0.0001449572,0.00009130605,0.001964585,0.0006481166,0.01450038],"genre_scores_gemma":[0.9080433,0.003277002,0.0739051,0.0008445795,0.0004705128,0.0004762704,0.002525411,0.0004814768,0.009976355],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01011592,"threshold_uncertainty_score":0.03384113,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03486758422580956,"score_gpt":0.2775667507472389,"score_spread":0.2426991665214294,"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."}}