PIPO-PITO bounded recurrent gain / bench_report.json
Failed on benchmark
1{
2 "bench_version": 1,
3 "track": "dynamics",
4 "model": "positive_rnn_fixed_gain_vs_pito",
5 "metric_direction": "lower is better",
6 "n_seeds": 8,
7 "baseline": {
8 "best_cfg": {
9 "lr": 0.01,
10 "gain": 1.5
11 },
12 "sweep": [
13 {
14 "cfg": {
15 "lr": 0.001,
16 "gain": 0.5
17 },
18 "mean": 0.007410152786178514
19 },
20 {
21 "cfg": {
22 "lr": 0.001,
23 "gain": 1.0
24 },
25 "mean": 0.004754966095788404
26 },
27 {
28 "cfg": {
29 "lr": 0.001,
30 "gain": 1.5
31 },
32 "mean": 0.0033777739736251533
33 },
34 {
35 "cfg": {
36 "lr": 0.003,
37 "gain": 0.5
38 },
39 "mean": 0.00141466197965201
40 },
41 {
42 "cfg": {
43 "lr": 0.003,
44 "gain": 1.0
45 },
46 "mean": 0.001106211231672205
47 },
48 {
49 "cfg": {
50 "lr": 0.003,
51 "gain": 1.5
52 },
53 "mean": 0.0007843969069654122
54 },
55 {
56 "cfg": {
57 "lr": 0.01,
58 "gain": 0.5
59 },
60 "mean": 0.0005034581408835948
61 },
62 {
63 "cfg": {
64 "lr": 0.01,
65 "gain": 1.0
66 },
67 "mean": 0.0003801305138040334
68 },
69 {
70 "cfg": {
71 "lr": 0.01,
72 "gain": 1.5
73 },
74 "mean": 0.00034718124516075477
75 }
76 ],
77 "full": {
78 "mean": 0.00034136135218432173,
79 "std": 0.00014732674795547322,
80 "per_seed": [
81 0.00023207295453175902,
82 0.00027615175349637866,
83 0.0003660226066131145,
84 0.0005144776660017669,
85 0.000629094778560102,
86 0.00019970460562035441,
87 0.00032621651189401746,
88 0.0001871499407570809
89 ],
90 "n": 8
91 }
92 },
93 "idea": {
94 "mean": 0.00037604584758810233,
95 "std": 9.077411363479533e-05,
96 "per_seed": [
97 0.00023790601699147373,
98 0.0003053620457649231,
99 0.0004130469460505992,
100 0.0005384435644373298,
101 0.00042402392136864364,
102 0.00027976761339232326,
103 0.0004256583342794329,
104 0.00038415833842009306
105 ],
106 "n": 8
107 },
108 "comparison": {
109 "delta_mean": 3.46844954037806e-05,
110 "idea_wins": 1,
111 "n_pairs": 8,
112 "per_seed_diffs": [
113 5.833062459714711e-06,
114 2.9210292268544436e-05,
115 4.702433943748474e-05,
116 2.396589843556285e-05,
117 -0.00020507085719145834,
118 8.006300777196884e-05,
119 9.944182238541543e-05,
120 0.00019700839766301215
121 ],
122 "p_value": 0.41415,
123 "mde": 9.544819469810646e-05,
124 "mde_rel_pct": 27.96104306692814,
125 "verdict": "no significant win",
126 "system_worked": false
127 },
128 "mechanism_signature": {
129 "prediction": {
130 "prediction": "decay slope = a*V when y>=V",
131 "predicted_rate": null,
132 "observed_rate": null,
133 "relative_error": null,
134 "n_traces": 0,
135 "confirmed": false
136 },
137 "selected_idea_cfg": {
138 "lr": 0.01,
139 "a": 1.0,
140 "b": 0.2,
141 "gain": 1.0
142 },
143 "matched_structure": "dynamics/control"
144 },
145 "idea_sweep": [
146 {
147 "cfg": {
148 "lr": 0.01,
149 "a": 1.0,
150 "b": 0.2,
151 "gain": 1.0
152 },
153 "result": {
154 "mean": 0.00037604584758810233,
155 "std": 9.077411363479533e-05,
156 "per_seed": [
157 0.00023790601699147373,
158 0.0003053620457649231,
159 0.0004130469460505992,
160 0.0005384435644373298,
161 0.00042402392136864364,
162 0.00027976761339232326,
163 0.0004256583342794329,
164 0.00038415833842009306
165 ],
166 "n": 8
167 },
168 "mechanism_signature": {
169 "prediction": "decay slope = a*V when y>=V",
170 "predicted_rate": null,
171 "observed_rate": null,
172 "relative_error": null,
173 "n_traces": 0,
174 "confirmed": false
175 }
176 },
177 {
178 "cfg": {
179 "lr": 0.001,
180 "a": 1.0,
181 "b": 0.2,
182 "gain": 1.0
183 },
184 "result": {
185 "mean": 0.004910496441880241,
186 "std": 0.0021499903596063027,
187 "per_seed": [
188 0.004891040734946728,
189 0.002062235027551651,
190 0.006821547634899616,
191 0.008928792551159859,
192 0.005618015769869089,
193 0.002876961138099432,
194 0.0052522034384310246,
195 0.002833175240084529
196 ],
197 "n": 8
198 },
199 "mechanism_signature": {
200 "prediction": "decay slope = a*V when y>=V",
201 "predicted_rate": 0.21370937302708626,
202 "observed_rate": 0.000306834834250262,
203 "relative_error": 0.9985642425041817,
204 "threshold_V": 0.21370937302708626,
205 "n_traces": 2,
206 "confirmed": false
207 }
208 },
209 {
210 "cfg": {
211 "lr": 0.01,
212 "a": 1.0,
213 "b": 0.2,
214 "gain": 1.0
215 },
216 "result": {
217 "mean": 0.00037604584758810233,
218 "std": 9.077411363479533e-05,
219 "per_seed": [
220 0.00023790601699147373,
221 0.0003053620457649231,
222 0.0004130469460505992,
223 0.0005384435644373298,
224 0.00042402392136864364,
225 0.00027976761339232326,
226 0.0004256583342794329,
227 0.00038415833842009306
228 ],
229 "n": 8
230 },
231 "mechanism_signature": {
232 "prediction": "decay slope = a*V when y>=V",
233 "predicted_rate": null,
234 "observed_rate": null,
235 "relative_error": null,
236 "n_traces": 0,
237 "confirmed": false
238 }
239 }
240 ],
241 "custom_track": null
242}