Wasserstein-Controlled Gaussian-Mixture Rollouts / results.json
Failed on benchmark
1{
2 "seed": 1729,
3 "parameters": {
4 "d": 0.9,
5 "q": 0.0025
6 },
7 "prediction_mode_separation": [
8 {
9 "d": 0.1,
10 "observed": 0.2,
11 "predicted": 0.2
12 },
13 {
14 "d": 0.3,
15 "observed": 0.6,
16 "predicted": 0.6
17 },
18 {
19 "d": 0.6,
20 "observed": 1.2,
21 "predicted": 1.2
22 },
23 {
24 "d": 0.9,
25 "observed": 1.8,
26 "predicted": 1.8
27 },
28 {
29 "d": 1.2,
30 "observed": 2.4,
31 "predicted": 2.4
32 }
33 ],
34 "prediction_merge_transition": [
35 {
36 "eps": 0.2,
37 "components": 2,
38 "pair_w2": 1.8,
39 "predicted_merge": false
40 },
41 {
42 "eps": 0.9,
43 "components": 2,
44 "pair_w2": 1.8,
45 "predicted_merge": false
46 },
47 {
48 "eps": 1.5,
49 "components": 2,
50 "pair_w2": 1.8,
51 "predicted_merge": false
52 },
53 {
54 "eps": 1.79,
55 "components": 2,
56 "pair_w2": 1.8,
57 "predicted_merge": false
58 },
59 {
60 "eps": 1.8,
61 "components": 1,
62 "pair_w2": 1.8,
63 "predicted_merge": true
64 },
65 {
66 "eps": 1.81,
67 "components": 1,
68 "pair_w2": 1.8,
69 "predicted_merge": true
70 },
71 {
72 "eps": 2.0,
73 "components": 1,
74 "pair_w2": 1.8,
75 "predicted_merge": true
76 }
77 ],
78 "prediction_chance_calibration": [
79 {
80 "d": 0.1,
81 "mixture": 0.9998542236739023,
82 "moment_gaussian": 0.9994427559174409,
83 "absolute_gap": 0.0004114677564613922
84 },
85 {
86 "d": 0.3,
87 "mixture": 0.9786298134890141,
88 "moment_gaussian": 0.9399062426562858,
89 "absolute_gap": 0.03872357083272837
90 },
91 {
92 "d": 0.6,
93 "mixture": 0.5974655681274085,
94 "moment_gaussian": 0.7933519853618862,
95 "absolute_gap": 0.19588641723447775
96 },
97 {
98 "d": 0.9,
99 "mixture": 0.5001457143252616,
100 "moment_gaussian": 0.7091744706768455,
101 "absolute_gap": 0.2090287563515839
102 },
103 {
104 "d": 1.2,
105 "mixture": 0.5000000004385546,
106 "moment_gaussian": 0.6608278736420903,
107 "absolute_gap": 0.16082787320353564
108 }
109 ],
110 "merge_error_sweep": [
111 {
112 "eps": 0.9,
113 "components": 2,
114 "sample_w2_to_unmerged": 0.024170828664971982,
115 "predicted_merged": false
116 },
117 {
118 "eps": 1.79,
119 "components": 2,
120 "sample_w2_to_unmerged": 0.08358656412767837,
121 "predicted_merged": false
122 },
123 {
124 "eps": 1.81,
125 "components": 1,
126 "sample_w2_to_unmerged": 0.4730228034266771,
127 "predicted_merged": true
128 },
129 {
130 "eps": 2.0,
131 "components": 1,
132 "sample_w2_to_unmerged": 0.4713694803125953,
133 "predicted_merged": true
134 }
135 ],
136 "rollout_comparison": [
137 {
138 "horizon": 1,
139 "components_exact": 2,
140 "components_gaussian": 1,
141 "sample_w2_error_gaussian_vs_mixture": 0.4726031866876991,
142 "exact_mean": 0.0,
143 "gaussian_mean": 0.0
144 },
145 {
146 "horizon": 2,
147 "components_exact": 4,
148 "components_gaussian": 1,
149 "sample_w2_error_gaussian_vs_mixture": 0.47105009625574146,
150 "exact_mean": 0.0,
151 "gaussian_mean": 0.0
152 },
153 {
154 "horizon": 4,
155 "components_exact": 16,
156 "components_gaussian": 1,
157 "sample_w2_error_gaussian_vs_mixture": 0.4759139455486349,
158 "exact_mean": 0.0,
159 "gaussian_mean": 0.0
160 },
161 {
162 "horizon": 6,
163 "components_exact": 64,
164 "components_gaussian": 1,
165 "sample_w2_error_gaussian_vs_mixture": 0.6536726020813202,
166 "exact_mean": 0.0,
167 "gaussian_mean": 0.0
168 },
169 {
170 "horizon": 8,
171 "components_exact": 256,
172 "components_gaussian": 1,
173 "sample_w2_error_gaussian_vs_mixture": 1.0517997749707413,
174 "exact_mean": 0.0,
175 "gaussian_mean": 0.0
176 }
177 ]
178}