{"id":"W4382468866","doi":"10.1609/aaai.v37i7.26061","title":"Adaptive Discrete Communication Bottlenecks with Dynamic Vector Quantization for Heterogeneous Representational Coarseness","year":2023,"lang":"en","type":"article","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","topic":"Advanced Memory and Neural Computing","field":"Engineering","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canadian Institute for Advanced Research; Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Discretization; Vector quantization; Codebook; Computer science; Bottleneck; Reinforcement learning; Quantization (signal processing); Hyperparameter; Artificial intelligence; Learning vector quantization; Robustness (evolution); Generalization; Theoretical computer science; Machine learning; Algorithm; Mathematics","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":[],"consensus_categories":[],"category_scores_codex":[0.0001447,0.0001392984,0.0001466055,0.00006963796,0.0001839558,0.00004553644,0.0003550311,0.00004396939,0.000007686367],"category_scores_gemma":[0.000101209,0.0001090989,0.00005219996,0.0003915016,0.0001224716,0.0001693396,0.00006084856,0.0001287591,0.00001157531],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003395791,"about_ca_system_score_gemma":0.00001704883,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000003382684,"about_ca_topic_score_gemma":0.00001414686,"domain_scores_codex":[0.99914,0.000008724915,0.0002700023,0.0002054463,0.0001860702,0.0001898054],"domain_scores_gemma":[0.9992557,0.000152804,0.0001330312,0.0001436539,0.000282485,0.00003228485],"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":[0.0005672399,0.00004768236,0.0001551604,0.0002004138,0.00006589363,7.266844e-7,0.001555403,0.5114432,0.2739845,0.1814867,0.00006099869,0.03043213],"study_design_scores_gemma":[0.00002715865,0.00008949541,0.0001325796,0.0001515913,0.00001056288,0.000002298672,0.0004604111,0.5523833,0.4272825,0.01933798,0.000008836883,0.0001133217],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.918736,0.00002796952,0.07879762,0.0004910883,0.0002343759,0.0007339919,0.00003038478,0.0003146184,0.0006339081],"genre_scores_gemma":[0.9985738,0.00004823539,0.00117804,0.00001696071,0.00002418413,0.00006130743,0.00001096532,0.00002444562,0.00006210489],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1621487,"threshold_uncertainty_score":0.4448925,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07642449509526475,"score_gpt":0.3147651972544703,"score_spread":0.2383407021592055,"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."}}