{"id":"W4292937774","doi":"10.17760/d20429166","title":"Improving instruction generation for vision-language navigation by reward designing","year":2021,"lang":"en","type":"dissertation","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Science North","funders":"","keywords":"Computer science; Generator (circuit theory); Function (biology); Artificial intelligence; Language model; Human–computer interaction","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":[0.0003497367,0.0002806085,0.0002440628,0.0001590189,0.0004331955,0.0005512889,0.0004947815,0.0003176553,0.00001924639],"category_scores_gemma":[0.0001856552,0.0003040393,0.0001243211,0.000394314,0.000009660532,0.0005779881,0.00005153027,0.0003385427,0.00001794254],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001432153,"about_ca_system_score_gemma":0.0002122336,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009009297,"about_ca_topic_score_gemma":0.00008326837,"domain_scores_codex":[0.9979647,0.00009568581,0.0004546864,0.0008943318,0.0003371225,0.000253425],"domain_scores_gemma":[0.9983781,0.0001136395,0.0004476703,0.000630377,0.0003490414,0.00008118857],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00000743564,0.00004233649,0.00001529021,0.000156591,0.00002293948,9.947513e-7,0.001791579,0.0006363365,0.6016032,0.002924403,0.002333577,0.3904653],"study_design_scores_gemma":[0.0004826096,0.00007988446,0.0002048133,0.000102359,0.00003516869,0.000008464828,0.0004960455,0.8690125,0.1281957,0.0002591333,0.0006429031,0.0004804333],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1315147,0.0002562295,0.8654847,0.0001671507,0.0008728321,0.0007012199,0.00001179791,0.000360761,0.0006305795],"genre_scores_gemma":[0.3294286,0.00001201373,0.6470047,0.0001146624,0.0004788755,0.0006982998,0.01556025,0.00007209939,0.006630449],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.8683762,"threshold_uncertainty_score":0.9999412,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.009269301080167717,"score_gpt":0.3029517608700136,"score_spread":0.2936824597898459,"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."}}