Exact Elasticity-Complex Message Passing / bench_report.json
Beats tuned baseline
1{
2 "bench_version": 1,
3 "track": "tetra_elasticity_complex",
4 "model": "mlp_tiny",
5 "metric_direction": "lower is better",
6 "n_seeds": 8,
7 "baseline": {
8 "best_cfg": {
9 "lr": 0.01
10 },
11 "sweep": [
12 {
13 "cfg": {
14 "lr": 0.001
15 },
16 "mean": 0.07423547469079494
17 },
18 {
19 "cfg": {
20 "lr": 0.003
21 },
22 "mean": 0.05538587085902691
23 },
24 {
25 "cfg": {
26 "lr": 0.01
27 },
28 "mean": 0.03730668965727091
29 }
30 ],
31 "full": {
32 "mean": 0.03292211447842419,
33 "std": 0.015248095167711763,
34 "per_seed": [
35 0.07105375826358795,
36 0.01817469112575054,
37 0.02902773581445217,
38 0.03097057342529297,
39 0.03421689197421074,
40 0.03229296952486038,
41 0.022874649614095688,
42 0.02476564608514309
43 ],
44 "n": 8
45 }
46 },
47 "idea": {
48 "mean": 0.00017560497508384287,
49 "std": 3.132081213307311e-05,
50 "per_seed": [
51 0.0002102894795825705,
52 0.00019068221445195377,
53 0.0001296501577598974,
54 0.00014417424972634763,
55 0.00018806532898452133,
56 0.00015338916273321956,
57 0.00016290618805214763,
58 0.00022568301938008517
59 ],
60 "n": 8
61 },
62 "comparison": {
63 "delta_mean": -0.03274650950334035,
64 "idea_wins": 8,
65 "n_pairs": 8,
66 "per_seed_diffs": [
67 -0.07084346878400538,
68 -0.017984008911298588,
69 -0.028898085656692274,
70 -0.03082639917556662,
71 -0.03402882664522622,
72 -0.03213958036212716,
73 -0.02271174342604354,
74 -0.024539963065763004
75 ],
76 "p_value": 0.0081,
77 "mde": 0.01362209922034768,
78 "mde_rel_pct": 41.37674458690995,
79 "verdict": "idea better (significant)",
80 "system_worked": true
81 },
82 "custom_track": {
83 "name": "tetra_elasticity_complex",
84 "file": "tetra_elasticity_complex.py",
85 "domain": "pde"
86 },
87 "algebra_sanity": {
88 "relative_D1D0": 0.0,
89 "relative_D2D1": 0.0,
90 "compatible_leak_exact_mean": 3.8892531210562334e-08,
91 "compatible_leak_unconstrained_mean": 2.4047226575300518
92 },
93 "idea_sweep": [
94 {
95 "cfg": {
96 "lr": 0.001
97 },
98 "result": {
99 "mean": 0.00052592100109905,
100 "std": 0.00016566725843712322,
101 "per_seed": [
102 0.000746735546272248,
103 0.0003983649949077517,
104 0.000376748270355165,
105 0.0003486424102447927,
106 0.0003762074338737875,
107 0.0005649435915984213,
108 0.0006035485421307385,
109 0.0007921772194094956
110 ],
111 "n": 8
112 }
113 },
114 {
115 "cfg": {
116 "lr": 0.003
117 },
118 "result": {
119 "mean": 0.00027460784986033104,
120 "std": 5.294603664377139e-05,
121 "per_seed": [
122 0.00034460300230421126,
123 0.00029504692065529525,
124 0.00018341776740271598,
125 0.0002037393132923171,
126 0.0002621814201120287,
127 0.0002868972660508007,
128 0.0002893827622756362,
129 0.00033159434678964317
130 ],
131 "n": 8
132 }
133 },
134 {
135 "cfg": {
136 "lr": 0.01
137 },
138 "result": {
139 "mean": 0.00017560497508384287,
140 "std": 3.132081213307311e-05,
141 "per_seed": [
142 0.0002102894795825705,
143 0.00019068221445195377,
144 0.0001296501577598974,
145 0.00014417424972634763,
146 0.00018806532898452133,
147 0.00015338916273321956,
148 0.00016290618805214763,
149 0.00022568301938008517
150 ],
151 "n": 8
152 }
153 }
154 ],
155 "protocol": {
156 "epochs": 35,
157 "batch": 128,
158 "grid": [
159 {
160 "lr": 0.001
161 },
162 {
163 "lr": 0.003
164 },
165 {
166 "lr": 0.01
167 }
168 ],
169 "paired_seeds": [
170 0,
171 1,
172 2,
173 3,
174 4,
175 5,
176 6,
177 7
178 ],
179 "structural_match": "PDE simplicial complex"
180 },
181 "mechanism_signature": {
182 "compatible_pred_abs_mean": 0.000795240506704431,
183 "observed_compatible_leakage": 3.8892531210562334e-08,
184 "unconstrained_observed_leakage": 2.4047226575300518,
185 "test_pred_observed_corr": 0.9987984258197132,
186 "test_pred_mean": 0.31432949006557465,
187 "test_observed_mean": 0.31512635573744774,
188 "confirmed": true
189 }
190}