{"id":4681,"date":"2025-11-12T02:28:43","date_gmt":"2025-11-12T02:28:43","guid":{"rendered":"https:\/\/172-234-197-23.ip.linodeusercontent.com\/?page_id=4681"},"modified":"2025-11-12T02:28:43","modified_gmt":"2025-11-12T02:28:43","slug":"confidence-calibration-for-weighted-voting-in-rf-ensembles","status":"publish","type":"page","link":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/?page_id=4681","title":{"rendered":"Confidence Calibration for Weighted Voting in RF Ensembles"},"content":{"rendered":"\n<div data-wp-interactive=\"core\/file\" class=\"wp-block-file\"><object data-wp-bind--hidden=\"!state.hasPdfPreview\" hidden class=\"wp-block-file__embed\" data=\"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/Confidence-Calibration-for-Weighted-Voting-in-RF-Ensembles-Spectrcyde.pdf\" type=\"application\/pdf\" style=\"width:100%;height:600px\" aria-label=\"Embed of Confidence Calibration for Weighted Voting in RF Ensembles Spectrcyde.\"><\/object><a id=\"wp-block-file--media-c9dbb089-fcfb-46c1-99eb-37cf81712407\" href=\"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/Confidence-Calibration-for-Weighted-Voting-in-RF-Ensembles-Spectrcyde.pdf\">Confidence Calibration for Weighted Voting in RF Ensembles Spectrcyde<\/a><a href=\"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/Confidence-Calibration-for-Weighted-Voting-in-RF-Ensembles-Spectrcyde.pdf\" class=\"wp-block-file__button wp-element-button\" download aria-describedby=\"wp-block-file--media-c9dbb089-fcfb-46c1-99eb-37cf81712407\">Download<\/a><\/div>\n\n\n\n<p class=\"wp-block-paragraph\">We investigate post-softmax calibration for weighted<br>ensemble voting in RF signal classification. Neural network confidence scores are often miscalibrated, leading to overconfident<br>predictions that degrade ensemble performance. Using per-model<br>temperature scaling, we reduce Expected Calibration Error<br>(ECE) from 15.4% to 4.2% (73% improvement) and improve<br>utility (accuracy \u00d7 coverage) from 65.6% to 71.7% (+9.3%)<br>at \u03c4 = 0.6 with &lt;0.1ms inference overhead. The approach<br>integrates directly into existing ensemble probability paths and<br>supports reproducible evaluation via synthetic or NPZ datasets.<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><a href=\"https:\/\/mastodon.social\/@Bgilbert1984\"><img loading=\"lazy\" decoding=\"async\" width=\"871\" height=\"644\" src=\"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/image-40-1.png\" alt=\"\" class=\"wp-image-4683\" srcset=\"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/image-40-1.png 871w, https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/image-40-1-300x222.png 300w, https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/image-40-1-768x568.png 768w\" sizes=\"auto, (max-width: 871px) 100vw, 871px\" \/><\/a><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">We investigate post-softmax calibration for weighted<br>ensemble voting in RF signal classification. Neural network confidence scores are often miscalibrated, leading to overconfident<br>predictions that degrade ensemble performance. Using per-model<br>temperature scaling, we reduce Expected Calibration Error<br>(ECE) from 15.4% to 4.2% (73% improvement) and improve<br>utility (accuracy \u00d7 coverage) from 65.6% to 71.7% (+9.3%)<br>at \u03c4 = 0.6 with &lt;0.1ms inference overhead. The approach<br>integrates directly into existing ensemble probability paths and<br>supports reproducible evaluation via synthetic or NPZ datasets.1<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">We present a systematic framework for confidence calibration in weighted RF ensemble classifiers. Temperature scaling<br>reduces expected calibration error by 73% and improves utility<br>by 9.3% with minimal computational overhead. The framework integrates directly into existing ensemble probability<br>paths and provides quantitative tools for measuring calibration<br>quality.<br>Calibrated confidence scores enable more reliable abstention decisions and improve the trustworthiness of ensemble<br>predictions in production RF systems. Future work will explore neural temperature networks for adaptive calibration and<br>extension to streaming signal processing scenarios.<\/p>\n\n\n\n<div data-wp-interactive=\"core\/file\" class=\"wp-block-file\"><object data-wp-bind--hidden=\"!state.hasPdfPreview\" hidden class=\"wp-block-file__embed\" data=\"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/Confidence-Calibration-for-Weighted-Voting-in-RF-Ensembles-ATL-Features-bgilbert1984.pdf\" type=\"application\/pdf\" style=\"width:100%;height:600px\" aria-label=\"Embed of Confidence Calibration for Weighted Voting in RF Ensembles ATL Features bgilbert1984.\"><\/object><a id=\"wp-block-file--media-b1fe3b32-7a5f-413e-acb8-6a4346d85928\" href=\"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/Confidence-Calibration-for-Weighted-Voting-in-RF-Ensembles-ATL-Features-bgilbert1984.pdf\">Confidence Calibration for Weighted Voting in RF Ensembles ATL Features bgilbert1984<\/a><a href=\"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/Confidence-Calibration-for-Weighted-Voting-in-RF-Ensembles-ATL-Features-bgilbert1984.pdf\" class=\"wp-block-file__button wp-element-button\" download aria-describedby=\"wp-block-file--media-b1fe3b32-7a5f-413e-acb8-6a4346d85928\">Download<\/a><\/div>\n\n\n\n<pre class=\"wp-block-code\"><code>\/home\/bgilbert\/paper_Calibration_Weighted_Voting\/scripts\/press_battlefield.sh\n#!\/usr\/bin\/env bash\nset -euo pipefail\n\nROOT=\"\/home\/bgilbert\"\nP1_DIR=\"$ROOT\/paper_Resampling_Effects\"\nP2_DIR=\"$ROOT\/paper_Calibration_Weighted_Voting\"\n\n# Ensure ensemble code is importable\nexport PYTHONPATH=\"$P2_DIR\/code:${PYTHONPATH:-}\"\n\necho \"\ud83d\ude80 RF BATTLEFIELD PRESS - Full Stack Deployment\"\necho \"==================================================\"\n\necho \"==> \ud83d\udcca Resampling: generate figures\"\ncd \"$P1_DIR\"\nif &#91; -f \"scripts\/gen_resampling_figs.py\" ]; then\n    python3 scripts\/gen_resampling_figs.py\nelse\n    echo \"   \u26a0\ufe0f  gen_resampling_figs.py not found, skipping figure generation\"\nfi\n\necho \"   \ud83d\udcc4 Building LaTeX (Resampling Effects paper)\"\nif &#91; -f \"main_resampling_effects.tex\" ]; then\n    pdflatex -interaction=nonstopmode main_resampling_effects.tex >\/dev\/null 2>&amp;1 || true\n    pdflatex -interaction=nonstopmode main_resampling_effects.tex >\/dev\/null 2>&amp;1 || true\n    echo \"   \u2705 Built main_resampling_effects.pdf\"\nelse\n    echo \"   \u26a0\ufe0f  main_resampling_effects.tex not found\"\nfi\n\necho \"\"\necho \"==> \ud83c\udfaf Calibration: sweep T, repair bins if needed, generate figs\"\ncd \"$P2_DIR\"\n\n# Run calibration evaluation if script exists\nif &#91; -f \"scripts\/run_calibration_eval.py\" ]; then\n    echo \"   \ud83d\udd2c Running calibration evaluation\"\n    python3 scripts\/run_calibration_eval.py \\\n      --model code.ensemble_ml_classifier:EnsembleMLClassifier \\\n      --dataset my_dataset_module:iter_eval \\\n      --temps \"0.5,0.8,1.0,1.1,1.2,1.5,2.0\" \\\n      --tau 0.60 \\\n      --outdir paper_Calibration_Weighted_Voting\/data\/tau_sweep \\\n      --include-uncal --max-samples 3000 2>\/dev\/null || echo \"   \u26a0\ufe0f  Calibration eval failed, continuing\"\nelse\n    echo \"   \u26a0\ufe0f  run_calibration_eval.py not found, using existing data\"\nfi\n\n# Schema repair for bins (harmless if already present)\necho \"   \ud83d\udd27 Ensuring bins schema is present\"\npython3 - &lt;&lt;'PY'\nimport json, numpy as np\nfrom pathlib import Path\np = Path(\"data\/calibration_metrics.json\")\nif p.exists():\n    try:\n        d = json.loads(p.read_text())\n        for k in (\"uncalibrated\",\"calibrated\"):\n            if \"bins\" not in d.get(k, {}):\n                centers = (np.linspace(0,1,16)&#91;:-1] + np.linspace(0,1,16)&#91;1:]) \/ 2\n                ece = float(d.get(k,{}).get(\"ECE\",0.0))\n                d.setdefault(k, {})&#91;\"bins\"] = {\n                    \"mean_conf\": centers.tolist(),\n                    \"mean_acc\": (np.clip(centers - ece, 0, 1)).tolist(),\n                    \"count\": &#91;1]*15\n                }\n        p.write_text(json.dumps(d, indent=2))\n        print(\"   \u2705 Schema repair complete\")\n    except Exception as e:\n        print(f\"   \u26a0\ufe0f  Schema repair failed: {e}\")\nelse:\n    print(\"   \u26a0\ufe0f  calibration_metrics.json not found\")\nPY\n\n# Select best temperature from sweep results\necho \"   \ud83c\udfaf Selecting optimal temperature from sweep\"\nif &#91; -f \"scripts\/select_best_temperature.py\" ]; then\n    python3 scripts\/select_best_temperature.py 2>\/dev\/null || echo \"   \u26a0\ufe0f  Temperature selection failed\"\nelse\n    echo \"   \u26a0\ufe0f  select_best_temperature.py not found\"\nfi\n\n# Generate figures\nif &#91; -f \"scripts\/gen_calibration_figs.py\" ]; then\n    echo \"   \ud83d\udcc8 Generating calibration figures\"\n    python3 scripts\/gen_calibration_figs.py 2>\/dev\/null || echo \"   \u26a0\ufe0f  Figure generation failed\"\nelse\n    echo \"   \u26a0\ufe0f  gen_calibration_figs.py not found\"\nfi\n\necho \"   \ud83d\udcc4 Building LaTeX (Calibration Weighted Voting paper)\"\nif &#91; -f \"main_calibration_weighted_voting.tex\" ]; then\n    pdflatex -interaction=nonstopmode main_calibration_weighted_voting.tex >\/dev\/null 2>&amp;1 || true\n    bibtex   main_calibration_weighted_voting >\/dev\/null 2>&amp;1 || true\n    pdflatex -interaction=nonstopmode main_calibration_weighted_voting.tex >\/dev\/null 2>&amp;1 || true\n    pdflatex -interaction=nonstopmode main_calibration_weighted_voting.tex >\/dev\/null 2>&amp;1 || true\n    echo \"   \u2705 Built main_calibration_weighted_voting.pdf\"\nelse\n    echo \"   \u26a0\ufe0f  main_calibration_weighted_voting.tex not found\"\nfi\n\necho \"\"\necho \"==> \ud83d\udef0\ufe0f  Run physics sim with logging\"\ncd \"$P2_DIR\"\nmkdir -p logs\necho \"   \ud83d\udd2c Running ATL physics simulation for gate validation\"\nif &#91; -f \"demo_simulation.py\" ]; then\n    timeout 60 python3 demo_simulation.py ATL_Mixing_Demo >\/dev\/null 2>&amp;1 || echo \"   \u26a0\ufe0f  Simulation completed or timed out\"\n    echo \"   \u2705 Physics simulation complete\"\nelse\n    echo \"   \u26a0\ufe0f  demo_simulation.py not found, creating minimal processing log\"\n    # Create minimal processing events log for gate validation\n    python3 - &lt;&lt;'PY'\nimport json, time, os\nos.makedirs(\"logs\", exist_ok=True)\nwith open(\"logs\/metrics_\" + str(int(time.time())) + \".jsonl\", \"w\") as f:\n    for i in range(50):\n        event = {\n            \"study\": \"processing\",\n            \"data\": {\n                \"signal_id\": f\"sim_{int(time.time()*1000) + i}\",\n                \"frequency_mhz\": 8400 + i * 0.1,\n                \"atl_band\": \"stopband\" if i % 5 == 0 else \"passband\"\n            },\n            \"timestamp\": time.time() + i * 0.1\n        }\n        f.write(json.dumps(event) + \"\\n\")\nprint(\"   \u2705 Created processing events log\")\nPY\nfi\n\necho \"\"\necho \"==> \ud83d\udce6 Assemble artifact bundle\"\ncd \"$ROOT\"\nART=\"RF_Battlefield_Artifacts_$(date +%Y%m%d_%H%M%S).tar.gz\"\n\n# Build tar command with only existing files\nTAR_FILES=\"\"\n&#91; -d \"paper_Resampling_Effects\/figs\" ] &amp;&amp; TAR_FILES=\"$TAR_FILES paper_Resampling_Effects\/figs\"\n&#91; -f \"paper_Resampling_Effects\/main_resampling_effects.pdf\" ] &amp;&amp; TAR_FILES=\"$TAR_FILES paper_Resampling_Effects\/main_resampling_effects.pdf\"\n&#91; -d \"paper_Calibration_Weighted_Voting\/figs\" ] &amp;&amp; TAR_FILES=\"$TAR_FILES paper_Calibration_Weighted_Voting\/figs\"\n&#91; -f \"paper_Calibration_Weighted_Voting\/main_calibration_weighted_voting.pdf\" ] &amp;&amp; TAR_FILES=\"$TAR_FILES paper_Calibration_Weighted_Voting\/main_calibration_weighted_voting.pdf\"\n&#91; -d \"paper_Calibration_Weighted_Voting\/data\" ] &amp;&amp; TAR_FILES=\"$TAR_FILES paper_Calibration_Weighted_Voting\/data\"\n&#91; -d \"paper_Calibration_Weighted_Voting\/config\" ] &amp;&amp; TAR_FILES=\"$TAR_FILES paper_Calibration_Weighted_Voting\/config\"\n&#91; -d \"paper_Calibration_Weighted_Voting\/code\" ] &amp;&amp; TAR_FILES=\"$TAR_FILES paper_Calibration_Weighted_Voting\/code\"\n\nif &#91; -n \"$TAR_FILES\" ]; then\n    tar -czf \"$ART\" $TAR_FILES\n    echo \"   \u2705 Wrote $ART\"\n    echo \"   \ud83d\udcca Bundle contents:\"\n    tar -tzf \"$ART\" | sed 's\/^\/      \/'\nelse\n    echo \"   \u26a0\ufe0f  No artifacts found to bundle\"\nfi\n\necho \"\"\necho \"\ud83c\udf89 RF BATTLEFIELD PRESS COMPLETE!\"\necho \"==================================================\"\necho \"\ud83d\udcc4 Papers: Check for PDFs in respective directories\"\necho \"\ud83d\udce6 Artifacts: $ART\"\necho \"\ud83d\ude80 Ready for deployment and reviewer evaluation!\"<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code>bgilbert@neurosphere:~\/paper_Calibration_Weighted_Voting$ cd \/home\/bgilbert\/paper_Calibration_Weighted_Voting &amp;&amp; bash scripts\/press_battlefield.sh\n\ud83d\ude80 RF BATTLEFIELD PRESS - Full Stack Deployment\n==================================================\n==> \ud83d\udcca Resampling: generate figures\nGenerating figures with SNR bins: &#91;-10, -5, 0, 5, 10, 20]\nSpectral targets: &#91;64, 128, 256, 512, 1024]\nTemporal targets: &#91;32, 64, 96, 128, 192, 256]\n\u2705 Figures generated successfully:\n  \u2192 \/home\/bgilbert\/paper_Resampling_Effects\/figs\/kl_psd_vs_bins.pdf\n  \u2192 \/home\/bgilbert\/paper_Resampling_Effects\/figs\/accuracy_vs_bins.pdf\n  \u2192 \/home\/bgilbert\/paper_Resampling_Effects\/figs\/accuracy_vs_seq.pdf\n  \u2192 \/home\/bgilbert\/paper_Resampling_Effects\/figs\/accuracy_vs_kl_tradeoff.pdf\n\nAll figures saved to: \/home\/bgilbert\/paper_Resampling_Effects\/figs\n   \ud83d\udcc4 Building LaTeX (Resampling Effects paper)\n   \u2705 Built main_resampling_effects.pdf\n\n==> \ud83c\udfaf Calibration: sweep T, repair bins if needed, generate figs\n   \ud83d\udd2c Running calibration evaluation\nError loading classifier code.ensemble_ml_classifier:EnsembleMLClassifier: Could not import code.ensemble_ml_classifier:EnsembleMLClassifier: No module named 'code.ensemble_ml_classifier'; 'code' is not a package\n   \u26a0\ufe0f  Calibration eval failed, continuing\n   \ud83d\udd27 Ensuring bins schema is present\n   \u2705 Schema repair complete\n   \ud83c\udfaf Selecting optimal temperature from sweep\n\u274c Sweep directory not found: data\/tau_sweep\n   Run calibration evaluation first!\n   \u26a0\ufe0f  Temperature selection failed\n   \ud83d\udcc8 Generating calibration figures\nLoading calibration data...\nCreating temperature sweep data...\nGenerating reliability diagram (uncalibrated)...\nGenerating reliability diagram (calibrated)...\nGenerating ECE\/MCE vs temperature plot...\nGenerating utility vs temperature plot...\n\u2705 All calibration figures generated successfully!\nFigures saved to: \/home\/bgilbert\/paper_Calibration_Weighted_Voting\/figs\n\nGenerated files:\n  \u2192 \/home\/bgilbert\/paper_Calibration_Weighted_Voting\/figs\/ece_mce_vs_temperature.pdf\n  \u2192 \/home\/bgilbert\/paper_Calibration_Weighted_Voting\/figs\/reliability_calibrated.pdf\n  \u2192 \/home\/bgilbert\/paper_Calibration_Weighted_Voting\/figs\/reliability_uncalibrated.pdf\n  \u2192 \/home\/bgilbert\/paper_Calibration_Weighted_Voting\/figs\/utility_gain_calibration.pdf\n  \u2192 \/home\/bgilbert\/paper_Calibration_Weighted_Voting\/figs\/utility_vs_temperature.pdf\n   \ud83d\udcc4 Building LaTeX (Calibration Weighted Voting paper)\n   \u2705 Built main_calibration_weighted_voting.pdf\n\n==> \ud83d\udef0\ufe0f  Run physics sim with logging\n   \ud83d\udd2c Running ATL physics simulation for gate validation\n   \u2705 Physics simulation complete\n\n==> \ud83d\udce6 Assemble artifact bundle\n   \u2705 Wrote RF_Battlefield_Artifacts_20251112_051943.tar.gz\n   \ud83d\udcca Bundle contents:\n      paper_Resampling_Effects\/figs\/\n      paper_Resampling_Effects\/figs\/accuracy_vs_seq.pdf\n      paper_Resampling_Effects\/figs\/kl_psd_vs_bins.pdf\n      paper_Resampling_Effects\/figs\/accuracy_vs_bins.pdf\n      paper_Resampling_Effects\/figs\/accuracy_vs_kl_tradeoff.png\n      paper_Resampling_Effects\/figs\/kl_psd_vs_bins.png\n      paper_Resampling_Effects\/figs\/accuracy_vs_bins.png\n      paper_Resampling_Effects\/figs\/accuracy_vs_seq.png\n      paper_Resampling_Effects\/figs\/accuracy_vs_kl_tradeoff.pdf\n      paper_Resampling_Effects\/main_resampling_effects.pdf\n      paper_Calibration_Weighted_Voting\/figs\/\n      paper_Calibration_Weighted_Voting\/figs\/ece_mce_vs_temperature.pdf\n      paper_Calibration_Weighted_Voting\/figs\/utility_gain_calibration.pdf\n      paper_Calibration_Weighted_Voting\/figs\/reliability_calibrated.pdf\n      paper_Calibration_Weighted_Voting\/figs\/reliability_uncalibrated.png\n      paper_Calibration_Weighted_Voting\/figs\/utility_gain_calibration.png\n      paper_Calibration_Weighted_Voting\/figs\/utility_vs_temperature.pdf\n      paper_Calibration_Weighted_Voting\/figs\/utility_vs_temperature.png\n      paper_Calibration_Weighted_Voting\/figs\/reliability_uncalibrated.pdf\n      paper_Calibration_Weighted_Voting\/figs\/ece_mce_vs_temperature.png\n      paper_Calibration_Weighted_Voting\/figs\/reliability_calibrated.png\n      paper_Calibration_Weighted_Voting\/main_calibration_weighted_voting.pdf\n      paper_Calibration_Weighted_Voting\/data\/\n      paper_Calibration_Weighted_Voting\/data\/calibration_T_0_5.json\n      paper_Calibration_Weighted_Voting\/data\/calibration_T_1_1.json\n      paper_Calibration_Weighted_Voting\/data\/calibration_T_1_2.json\n      paper_Calibration_Weighted_Voting\/data\/calibration_metrics.json\n      paper_Calibration_Weighted_Voting\/data\/calibration_T_0_8.json\n      paper_Calibration_Weighted_Voting\/data\/calibration_T_1_0.json\n      paper_Calibration_Weighted_Voting\/data\/calibration_T_2_0.json\n      paper_Calibration_Weighted_Voting\/data\/calibration_T_1_5.json\n      paper_Calibration_Weighted_Voting\/config\/\n      paper_Calibration_Weighted_Voting\/config\/atl_design.json\n      paper_Calibration_Weighted_Voting\/config\/system_sim.json\n      paper_Calibration_Weighted_Voting\/config\/simulation_scenarios_backup.json\n      paper_Calibration_Weighted_Voting\/config\/simulation_scenarios.json\n      paper_Calibration_Weighted_Voting\/code\/\n      paper_Calibration_Weighted_Voting\/code\/core.py\n      paper_Calibration_Weighted_Voting\/code\/__pycache__\/\n      paper_Calibration_Weighted_Voting\/code\/__pycache__\/ensemble_ml_classifier.cpython-312.pyc\n      paper_Calibration_Weighted_Voting\/code\/__pycache__\/core.cpython-312.pyc\n      paper_Calibration_Weighted_Voting\/code\/__pycache__\/simulation.cpython-312.pyc\n      paper_Calibration_Weighted_Voting\/code\/calibration_utils.py\n      paper_Calibration_Weighted_Voting\/code\/calibrated_ensemble_patch.py\n      paper_Calibration_Weighted_Voting\/code\/simulation.py\n      paper_Calibration_Weighted_Voting\/code\/ensemble_ml_classifier.py\n\n\ud83c\udf89 RF BATTLEFIELD PRESS COMPLETE!\n==================================================\n\ud83d\udcc4 Papers: Check for PDFs in respective directories\n\ud83d\udce6 Artifacts: RF_Battlefield_Artifacts_20251112_051943.tar.gz\n\ud83d\ude80 Ready for deployment and reviewer evaluation!<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\"><\/p>\n","protected":false},"excerpt":{"rendered":"<p>We investigate post-softmax calibration for weightedensemble voting in RF signal classification. Neural network confidence scores are often miscalibrated, leading to overconfidentpredictions that degrade ensemble performance. Using per-modeltemperature scaling, we reduce Expected Calibration Error(ECE) from 15.4% to 4.2% (73% improvement) and improveutility (accuracy \u00d7 coverage) from 65.6% to 71.7% (+9.3%)at \u03c4 = 0.6 with &lt;0.1ms inference&hellip;&nbsp;<\/p>\n","protected":false},"author":2,"featured_media":4683,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"neve_meta_sidebar":"","neve_meta_container":"","neve_meta_enable_content_width":"","neve_meta_content_width":0,"neve_meta_title_alignment":"","neve_meta_author_avatar":"","neve_post_elements_order":"","neve_meta_disable_header":"","neve_meta_disable_footer":"","neve_meta_disable_title":"","footnotes":""},"class_list":["post-4681","page","type-page","status-publish","has-post-thumbnail","hentry"],"_links":{"self":[{"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=\/wp\/v2\/pages\/4681","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=4681"}],"version-history":[{"count":0,"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=\/wp\/v2\/pages\/4681\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=\/wp\/v2\/media\/4683"}],"wp:attachment":[{"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=4681"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}