{ "bench_version": 1, "track": "anisotropic_oblique_bvp", "model": "mlp_tiny", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.003, "weight_decay": 0.0 }, "sweep": [ { "cfg": { "lr": 0.001, "weight_decay": 0.0 }, "mean": 0.00012391583004500717 }, { "cfg": { "lr": 0.003, "weight_decay": 0.0 }, "mean": 0.00010377542184869526 }, { "cfg": { "lr": 0.006, "weight_decay": 0.0 }, "mean": 0.00011149384499731241 } ], "full": { "mean": 9.435732454221579e-05, "std": 2.4013090964758765e-05, "per_seed": [ 0.00012657140905503184, 0.00012852151121478528, 5.6132030294975266e-05, 0.00010387673682998866, 8.006240386748686e-05, 9.653119195718318e-05, 9.42463957471773e-05, 6.891691737109795e-05 ], "n": 8 } }, "idea": { "mean": 0.00011699842434609309, "std": 3.095430559427717e-05, "per_seed": [ 0.00016582579701207578, 0.0001437541941413656, 6.513581320177764e-05, 9.932620014296845e-05, 0.00011056653602281585, 0.0001362798793707043, 0.00012959061132278293, 8.550836355425417e-05 ], "n": 8, "cfg": { "lr": 0.003, "weight_decay": 0.0, "bc_weight": 0.1 }, "selected_cfg": { "lr": 0.003, "weight_decay": 0.0, "bc_weight": 0.1 } }, "comparison": { "delta_mean": 2.2641099803877296e-05, "idea_wins": 1, "n_pairs": 8, "per_seed_diffs": [ 3.9254387957043946e-05, 1.523268292658031e-05, 9.003782906802371e-06, -4.550536687020212e-06, 3.050413215532899e-05, 3.974868741352111e-05, 3.534421557560563e-05, 1.6591446183156222e-05 ], "p_value": 0.0152, "mde": 1.3440697763296005e-05, "mde_rel_pct": 14.244466795243426, "verdict": "idea worse (significant)", "system_worked": false }, "mechanism_signature": { "mechanism_signature": { "prediction": "boundary-aware PDE training reduces measured oblique/no-flux residuals", "mean_baseline_oblique_residual": 0.02720322646200657, "mean_idea_oblique_residual": 0.01003198220860213, "mean_baseline_normal_residual": 0.04276570677757263, "mean_idea_normal_residual": 0.008119827311020344, "per_seed": [ { "seed": 0, "baseline_oblique_abs_derivative": 0.03964276611804962, "idea_oblique_abs_derivative": 0.01283207256346941, "baseline_normal_abs_derivative": 0.03705023229122162, "idea_normal_abs_derivative": 0.009148340672254562 }, { "seed": 1, "baseline_oblique_abs_derivative": 0.020765259861946106, "idea_oblique_abs_derivative": 0.008115937002003193, "baseline_normal_abs_derivative": 0.04100044071674347, "idea_normal_abs_derivative": 0.010081736370921135 }, { "seed": 2, "baseline_oblique_abs_derivative": 0.020734591409564018, "idea_oblique_abs_derivative": 0.009465154260396957, "baseline_normal_abs_derivative": 0.06341361999511719, "idea_normal_abs_derivative": 0.01047296728938818 }, { "seed": 3, "baseline_oblique_abs_derivative": 0.023547939956188202, "idea_oblique_abs_derivative": 0.009573541581630707, "baseline_normal_abs_derivative": 0.027527321130037308, "idea_normal_abs_derivative": 0.00741744227707386 }, { "seed": 4, "baseline_oblique_abs_derivative": 0.02001587674021721, "idea_oblique_abs_derivative": 0.009027393534779549, "baseline_normal_abs_derivative": 0.04191877692937851, "idea_normal_abs_derivative": 0.008082550019025803 }, { "seed": 5, "baseline_oblique_abs_derivative": 0.026504037901759148, "idea_oblique_abs_derivative": 0.007130592130124569, "baseline_normal_abs_derivative": 0.031208854168653488, "idea_normal_abs_derivative": 0.007085439749062061 }, { "seed": 6, "baseline_oblique_abs_derivative": 0.0324188694357872, "idea_oblique_abs_derivative": 0.016259243711829185, "baseline_normal_abs_derivative": 0.03935849666595459, "idea_normal_abs_derivative": 0.007526469882577658 }, { "seed": 7, "baseline_oblique_abs_derivative": 0.033996470272541046, "idea_oblique_abs_derivative": 0.007851922884583473, "baseline_normal_abs_derivative": 0.060647912323474884, "idea_normal_abs_derivative": 0.005143672227859497 } ], "confirmed": true }, "custom_track": { "name": "anisotropic_oblique_bvp", "file": "pde_boundary_track.py", "domain": "pde" }, "idea_sweep": [ { "cfg": { "lr": 0.003, "weight_decay": 0.0, "bc_weight": 0.1 }, "mean": 0.00011699842434609309, "std": 3.095430559427717e-05, "per_seed": [ 0.00016582579701207578, 0.0001437541941413656, 6.513581320177764e-05, 9.932620014296845e-05, 0.00011056653602281585, 0.0001362798793707043, 0.00012959061132278293, 8.550836355425417e-05 ] }, { "cfg": { "lr": 0.003, "weight_decay": 0.0, "bc_weight": 1.0 }, "mean": 0.00016314282402163371, "std": 5.0713334733415886e-05, "per_seed": [ 0.00013824015331920236, 0.0002775485045276582, 9.841256542131305e-05, 0.00013486476382240653, 0.00015113064728211612, 0.00019398487347643822, 0.00017369228589814156, 0.00013726879842579365 ] }, { "cfg": { "lr": 0.003, "weight_decay": 0.0, "bc_weight": 5.0 }, "mean": 0.00022120710127637722, "std": 5.5719673809869554e-05, "per_seed": [ 0.00020971496996935457, 0.0002884801651816815, 0.00013389925879891962, 0.0002538147964514792, 0.00018812052439898252, 0.00022245440050028265, 0.0003070295206271112, 0.00016614317428320646 ] } ] }, "protocol_note": "Custom PDE track required: no built-in track contains boundary-value/PDE structure." }