{ "bench_version": 1, "track": "poisson_boundary", "model": "smooth_mlp_shared", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "sweep": [ { "cfg": { "lr": 0.001 }, "mean": 0.09443562291562557 }, { "cfg": { "lr": 0.003 }, "mean": 0.026866592932492495 }, { "cfg": { "lr": 0.01 }, "mean": 0.021620961604639888 } ], "best_cfg": { "lr": 0.01 }, "full": { "config": { "lr": 0.01, "pde_weight": 0.0 }, "per_seed": [ 0.027827076613903046, 0.013140390627086163, 0.02447698265314102, 0.02103939652442932, 0.019167248159646988, 0.014232118614017963, 0.017978858202695847, 0.019848095253109932 ], "details": [ { "metric": 0.027827076613903046, "ramp_epochs": [], "early_pred": 0.17212451994419098, "late_pred": 0.07173162906923715, "early_pde": 0.016137513642509777, "late_pde": 57.825637928758034, "output_var": 0.16997186839580536 }, { "metric": 0.013140390627086163, "ramp_epochs": [], "early_pred": 0.14410895109176636, "late_pred": 0.06673828408341198, "early_pde": 0.016183204328020413, "late_pde": 49.783568495610616, "output_var": 0.14312273263931274 }, { "metric": 0.02447698265314102, "ramp_epochs": [], "early_pred": 0.2193710853656133, "late_pred": 0.06691976044984425, "early_pde": 0.033302336310346924, "late_pde": 44.65269423695281, "output_var": 0.15733948349952698 }, { "metric": 0.02103939652442932, "ramp_epochs": [], "early_pred": 0.20242826143900552, "late_pred": 0.06018218672012582, "early_pde": 0.03576013880471388, "late_pde": 46.459858779740685, "output_var": 0.20421859622001648 }, { "metric": 0.019167248159646988, "ramp_epochs": [], "early_pred": 0.19904146095116934, "late_pred": 0.062198137974037844, "early_pde": 0.030983473853363346, "late_pde": 49.05369365796008, "output_var": 0.20605206489562988 }, { "metric": 0.014232118614017963, "ramp_epochs": [], "early_pred": 0.15507098038991293, "late_pred": 0.06462596554089994, "early_pde": 0.01172618723164002, "late_pde": 23.61071216627298, "output_var": 0.171492800116539 }, { "metric": 0.017978858202695847, "ramp_epochs": [], "early_pred": 0.19897286593914032, "late_pred": 0.07070110004176111, "early_pde": 0.01596340723335743, "late_pde": 40.135394351508964, "output_var": 0.15762735903263092 }, { "metric": 0.019848095253109932, "ramp_epochs": [], "early_pred": 0.12244720011949539, "late_pred": 0.06269879924023852, "early_pde": 0.006550400673101346, "late_pde": 51.49694557055174, "output_var": 0.18547753989696503 } ], "mean": 0.019713770831003785, "std": 0.004566942238936102 } }, "idea": { "config": { "lr": 0.003, "pde_weight": 1.0 }, "per_seed": [ 0.10768520832061768, 0.07102982699871063, 0.07831890136003494, 0.08951634913682938, 0.08078285306692123, 0.07756161689758301, 0.10001540929079056, 0.08986665308475494 ], "details": [ { "metric": 0.10768520832061768, "ramp_epochs": [ 15, 15 ], "early_pred": 0.14781861007213593, "late_pred": 0.09978784752242706, "early_pde": 0.005272098972151677, "late_pde": 0.6819872995290686, "output_var": 0.0017096633091568947 }, { "metric": 0.07102982699871063, "ramp_epochs": [ 12 ], "early_pred": 0.1365677664677302, "late_pred": 0.09450832053142436, "early_pde": 0.0048792896171410876, "late_pde": 0.15780467033276663, "output_var": 0.01197989284992218 }, { "metric": 0.07831890136003494, "ramp_epochs": [ 13 ], "early_pred": 0.11298269033432007, "late_pred": 0.08290284718660747, "early_pde": 0.0038497335044667125, "late_pde": 0.6933216733912773, "output_var": 0.006689825095236301 }, { "metric": 0.08951634913682938, "ramp_epochs": [ 13 ], "early_pred": 0.13521990676720938, "late_pred": 0.08473961953731145, "early_pde": 0.007382349111139774, "late_pde": 2.1494412894415507, "output_var": 0.004558507353067398 }, { "metric": 0.08078285306692123, "ramp_epochs": [ 11 ], "early_pred": 0.14865803718566895, "late_pred": 0.09791501523817286, "early_pde": 0.007154937057445447, "late_pde": 0.5216962382854784, "output_var": 0.011441169306635857 }, { "metric": 0.07756161689758301, "ramp_epochs": [ 15, 19 ], "early_pred": 0.13096309204896292, "late_pred": 0.09107760997379527, "early_pde": 0.003719212933598707, "late_pde": 0.5973782197722945, "output_var": 0.007852709852159023 }, { "metric": 0.10001540929079056, "ramp_epochs": [ 14, 14 ], "early_pred": 0.12350016832351685, "late_pred": 0.08857208950554624, "early_pde": 0.007426746034373839, "late_pde": 0.808233292773366, "output_var": 0.0023250302765518427 }, { "metric": 0.08986665308475494, "ramp_epochs": [ 12, 13 ], "early_pred": 0.11297053098678589, "late_pred": 0.08425354891840149, "early_pde": 0.005976979232703646, "late_pde": 0.4342757718199316, "output_var": 0.002663834486156702 } ], "mean": 0.0868471022695303, "std": 0.011563873184926184 }, "comparison": { "delta_mean": 0.06713333143852651, "idea_wins": 0, "n_pairs": 8, "per_seed_diffs": [ 0.07985813170671463, 0.05788943637162447, 0.05384191870689392, 0.06847695261240005, 0.061615604907274246, 0.06332949828356504, 0.08203655108809471, 0.07001855783164501 ], "p_value": 0.0081, "mde": 0.008368594524021243, "mde_rel_pct": 42.45050120426469, "verdict": "idea worse (significant)", "system_worked": false }, "mechanism_signature": { "prediction": "predictive loss stabilizes before strong PDE reduction", "early_predictive_loss": 0.1310851002732913, "late_predictive_loss": 0.09046961230171077, "early_pde_residual": 0.005707668307877611, "late_pde_residual": 0.7555173069182167, "confirmed": false, "math_check_max_abs_laplacian": 0.0 }, "custom_track": { "name": "poisson_boundary", "file": "/home/maxwelhelp/all/math2nn/bench/custom_tracks/poisson_boundary.py", "domain": "pde" } }