{ "bench_version": 1, "track": "obb_clearance_trajectory", "model": "mlp_tiny", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.01, "epochs": 24 }, "sweep": [ { "cfg": { "lr": 0.001, "epochs": 24 }, "mean": 2.5900412797927856 }, { "cfg": { "lr": 0.003, "epochs": 24 }, "mean": 1.0762107968330383 }, { "cfg": { "lr": 0.01, "epochs": 24 }, "mean": 0.08821548242121935 } ], "full": { "mean": 0.11757882172241807, "std": 0.04988984373475929, "per_seed": [ 0.11706195771694183, 0.1304021030664444, 0.03879903629422188, 0.06659883260726929, 0.1332305371761322, 0.21891747415065765, 0.13242962956428528, 0.10319100320339203 ], "n": 8 } }, "idea": { "mean": 0.1973940273746848, "std": 0.1311616228648052, "per_seed": [ 0.126096710562706, 0.13761737942695618, 0.15668733417987823, 0.12452193349599838, 0.203719824552536, 0.5375758409500122, 0.12353026866912842, 0.16940292716026306 ], "n": 8, "signature_values": [ [ 0.22611111402511597, -0.5267328023910522, 0.3687500059604645 ], [ 0.14370578527450562, -0.4949674904346466, 0.37312498688697815 ], [ 0.3007834553718567, -0.5578157901763916, 0.27250000834465027 ], [ 0.20459963381290436, -0.4704010784626007, 0.43437498807907104 ], [ 0.22297126054763794, -0.26175621151924133, 0.17687499523162842 ], [ 0.6471734642982483, -0.03403729200363159, 0.0012499999720603228 ], [ 0.1714925915002823, -0.5904415249824524, 0.4350000023841858 ], [ 0.2024228423833847, -0.5173735618591309, 0.35874998569488525 ] ] }, "comparison": { "delta_mean": 0.07981520565226674, "idea_wins": 1, "n_pairs": 8, "per_seed_diffs": [ 0.00903475284576416, 0.00721527636051178, 0.11788829788565636, 0.057923100888729095, 0.07048928737640381, 0.31865836679935455, -0.00889936089515686, 0.06621192395687103 ], "p_value": 0.02415, "mde": 0.08792278503854178, "mde_rel_pct": 74.77773952022693, "verdict": "idea worse (significant)", "system_worked": false }, "mechanism_signature": { "math_check": { "predicted_exact_zero_m": 1.1, "observed_exact_zero_m": 1.100000023841858, "observed_smooth_zero_m": 1.0299999713897705, "smoothing_bound": [ { "tau": 0.2, "error": 0.003949284553527832, "bound": 0.2772588722239781 }, { "tau": 0.1, "error": 3.3736228942871094e-05, "bound": 0.13862943611198905 }, { "tau": 0.05, "error": 0.0, "bound": 0.06931471805599453 } ], "active_axis_gradient_dx": -0.0, "passed": true }, "idea_sweep": [ { "cfg": { "lr": 0.001, "epochs": 24, "lambda_barrier": 0.08, "tau": 0.1, "m0": 0.05, "beta": 0.05 }, "mean": 2.590525209903717 }, { "cfg": { "lr": 0.003, "epochs": 24, "lambda_barrier": 0.08, "tau": 0.1, "m0": 0.05, "beta": 0.05 }, "mean": 1.1768128275871277 }, { "cfg": { "lr": 0.01, "epochs": 24, "lambda_barrier": 0.08, "tau": 0.1, "m0": 0.05, "beta": 0.05 }, "mean": 0.1362308394163847 } ], "predicted_vs_observed": { "predicted_smoothed_margin_mean": 0.264907518401742, "observed_exact_sat_margin_mean": -0.4316907189786434, "predicted_collision_fraction": 0.30257812156924047 }, "prediction": "barrier should reduce collision fraction and move margins upward", "confirmed": true }, "custom_track": { "name": "obb_clearance_trajectory", "file": "obb_track.py", "domain": "geometric_collision_avoidance" } }