Lie-Poisson Hamiltonian latent block / bench_report.json
Failed on benchmark
1{
2 "bench_version": 1,
3 "track": "dynamics",
4 "model": "rnn_small",
5 "metric_direction": "lower is better",
6 "n_seeds": 8,
7 "baseline": {
8 "best_cfg": {
9 "lr": 0.003
10 },
11 "sweep": [
12 {
13 "cfg": {
14 "lr": 0.001
15 },
16 "mean": 0.009041695157065988
17 },
18 {
19 "cfg": {
20 "lr": 0.003
21 },
22 "mean": 0.004719233373180032
23 },
24 {
25 "cfg": {
26 "lr": 0.01
27 },
28 "mean": 0.006545157753862441
29 }
30 ],
31 "full": {
32 "mean": 0.005879616655874997,
33 "std": 0.0028056293033459226,
34 "per_seed": [
35 0.0026779656764119864,
36 0.004133712034672499,
37 0.008910770528018475,
38 0.0031544852536171675,
39 0.006136221811175346,
40 0.0032570217736065388,
41 0.00849937554448843,
42 0.010267380625009537
43 ],
44 "n": 8
45 }
46 },
47 "idea": {
48 "mean": 0.02712346729822457,
49 "std": 0.004941447293628062,
50 "per_seed": [
51 0.028157928958535194,
52 0.02860170602798462,
53 0.0332900770008564,
54 0.03025132603943348,
55 0.018054386600852013,
56 0.02373768761754036,
57 0.02242955192923546,
58 0.032465074211359024
59 ],
60 "n": 8
61 },
62 "comparison": {
63 "delta_mean": 0.02124385064234957,
64 "idea_wins": 0,
65 "n_pairs": 8,
66 "per_seed_diffs": [
67 0.025479963282123208,
68 0.02446799399331212,
69 0.024379306472837925,
70 0.027096840785816312,
71 0.011918164789676666,
72 0.02048066584393382,
73 0.013930176384747028,
74 0.022197693586349487
75 ],
76 "p_value": 0.0081,
77 "mde": 0.004627304642862352,
78 "mde_rel_pct": 78.70078805628737,
79 "verdict": "idea worse (significant)",
80 "system_worked": false
81 },
82 "mechanism_signature": {
83 "mechanism_signature": {
84 "trained_model": "LiePoissonRNN",
85 "scales": [
86 0.25,
87 0.5,
88 1.0
89 ],
90 "observed_rms_effect": [
91 0.007525605789851397,
92 0.014394661411643028,
93 0.025677648605778813
94 ],
95 "observed_loglog_exponent": 0.8853167356267617,
96 "confirmed": false
97 },
98 "idea_sweep": [
99 {
100 "cfg": {
101 "lr": 0.001
102 },
103 "eval": {
104 "mean": 0.20517639257013798,
105 "std": 0.05351073177265944,
106 "per_seed": [
107 0.16992242634296417,
108 0.199637770652771,
109 0.268087774515152,
110 0.29078808426856995,
111 0.18213585019111633,
112 0.2410622090101242,
113 0.16997386515140533,
114 0.11980316042900085
115 ],
116 "n": 8
117 }
118 },
119 {
120 "cfg": {
121 "lr": 0.003
122 },
123 "eval": {
124 "mean": 0.03456594352610409,
125 "std": 0.009361537439374885,
126 "per_seed": [
127 0.02949955314397812,
128 0.028048669919371605,
129 0.052656497806310654,
130 0.03622344508767128,
131 0.042653679847717285,
132 0.03863559663295746,
133 0.021713633090257645,
134 0.027096472680568695
135 ],
136 "n": 8
137 }
138 },
139 {
140 "cfg": {
141 "lr": 0.01
142 },
143 "eval": {
144 "mean": 0.02712346729822457,
145 "std": 0.004941447293628062,
146 "per_seed": [
147 0.028157928958535194,
148 0.02860170602798462,
149 0.0332900770008564,
150 0.03025132603943348,
151 0.018054386600852013,
152 0.02373768761754036,
153 0.02242955192923546,
154 0.032465074211359024
155 ],
156 "n": 8
157 }
158 }
159 ],
160 "parameter_count_baseline": 22261,
161 "parameter_count_idea": 6254
162 }
163}