{"id":"W4380558639","doi":"10.48550/arxiv.2306.06968","title":"Can Forward Gradient Match Backpropagation?","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute","funders":"Natural Sciences and Engineering Research Council of Canada; Sorbonne Université; Grand Équipement National De Calcul Intensif; Agence Nationale de la Recherche; Compute Canada","keywords":"Backpropagation; Artificial neural network; Computer science; Computation; Isotropy; Gradient descent; Noise (video); Gradient method; Artificial intelligence; Algorithm; Physics; Image (mathematics); Optics","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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0001301749,0.0002997181,0.0002687641,0.0002214718,0.0002377194,0.0001014276,0.002227886,0.0001812905,0.00001121484],"category_scores_gemma":[0.00001851459,0.0003573867,0.0001796488,0.001162523,0.00009172087,0.0002357116,0.002582212,0.0005162496,0.0004512475],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002976657,"about_ca_system_score_gemma":0.0001335651,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001061038,"about_ca_topic_score_gemma":0.0001842292,"domain_scores_codex":[0.9978396,0.00007250727,0.0002182757,0.00132505,0.0001153174,0.0004291933],"domain_scores_gemma":[0.9974578,0.0001218283,0.0002535414,0.001803365,0.0001509653,0.0002125174],"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":[0.000005041832,0.00004445137,0.0008428032,0.00003871481,0.00004586405,0.0001378678,0.0001360117,0.4820262,0.00002455333,0.5134572,0.002370479,0.0008708632],"study_design_scores_gemma":[0.0002067814,0.00002539881,0.001673667,0.00004790423,0.00003427715,0.00000444277,0.00003140595,0.4898991,0.0001360058,0.505846,0.001627441,0.0004675453],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04709387,0.00001758397,0.9476396,0.00193813,0.0007458372,0.0005917543,0.00003188184,0.001042049,0.0008993271],"genre_scores_gemma":[0.9873567,0.0001846802,0.006709239,0.0001536375,0.000107065,0.00001004723,0.00004952477,0.00003495536,0.005394138],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9409303,"threshold_uncertainty_score":0.9998878,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0851859728657592,"score_gpt":0.2083090153710565,"score_spread":0.1231230425052973,"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."}}