{"id":"W4226316739","doi":"10.2196/34834","title":"Pretrained Transformer Language Models Versus Pretrained Word Embeddings for the Detection of Accurate Health Information on Arabic Social Media: Comparative Study","year":2022,"lang":"en","type":"article","venue":"JMIR Formative Research","topic":"Topic Modeling","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Taibah University; Science Foundation Ireland","keywords":"Computer science; Social media; Natural language processing; Leverage (statistics); Language model; Artificial intelligence; Transformer; Arabic; Health informatics; Machine learning; Information retrieval; World Wide Web; Linguistics; Medicine; Public health","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"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":[0.003035977,0.002778107,0.0009073067,0.001119575,0.000403738,0.001237705,0.001232014,0.00136249,0.002324819],"category_scores_gemma":[0.011332,0.0005526656,0.0009721706,0.0007454444,0.0006451747,0.003043144,0.001250368,0.002902193,0.001906551],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001168322,"about_ca_system_score_gemma":0.001134613,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01516035,"about_ca_topic_score_gemma":0.0134535,"domain_scores_codex":[0.9985689,0.0005744651,0.0001350461,0.0003719684,0.0001846621,0.000165014],"domain_scores_gemma":[0.9937354,0.004218644,0.0002493977,0.0004701987,0.001162116,0.0001642883],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002892516,0.002335179,0.04981293,0.001344673,0.0009617928,0.0007785661,0.0008212368,0.3310911,0.0099249,0.001521854,0.01586967,0.5826455],"study_design_scores_gemma":[0.00005430739,0.0005556446,0.00552586,0.00009201175,0.0001580391,0.0001364267,0.0004071522,0.9841897,0.006519585,0.001027296,0.0012844,0.00004958406],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9341162,0.00631552,0.04774769,0.0009230722,0.0006409636,0.0002487116,0.00190298,0.002341219,0.005763644],"genre_scores_gemma":[0.9662161,0.001205147,0.02457559,0.0003096467,0.00009656853,0.0001324131,0.004847942,0.0001335334,0.002483146],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01516035,"threshold_uncertainty_score":0.03014421,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1444257212986051,"score_gpt":0.4327990415068596,"score_spread":0.2883733202082545,"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."}}