| … | … | @@ -13,9 +13,11 @@ The Rocca agent is able to learn and plan the desired cognitive |
| 13 | 13 | schematics leading to getting/accumulating a reward via pattern mining |
| 14 | 14 | and temporal deduction. |
| 15 | 15 | |
| 16 | | The three optimal action plans required to get a reward are: |
| 16 | The following are optimal action plans the agent learned to get a reward: |
| 17 | 17 | |
| 18 | | 1. `outside(self, house) ∧ do(go_to_key) ↝ hold(self, key)` |
| 18 | 1. Monoaction plans: |
| 19 | |
| 20 | 1.1. `outside(self, house) ∧ do(go_to_key) ↝ hold(self, key)` |
| 19 | 21 | |
| 20 | 22 | ```scheme |
| 21 | 23 | (BackPredictiveImplicationScopeLink (stv 1 0.0111248) |
| … | … | @@ -46,7 +48,7 @@ The three optimal action plans required to get a reward are: |
| 46 | 48 | ) ; [e6a8f21e6b37d8f0][1] |
| 47 | 49 | ) ; [ff069058911f0233][1] |
| 48 | 50 | ``` |
| 49 | | 2. `outside(self, house) ∧ hold(self, key) ∧ do(go_to_house) ↝ inside(self, house)` |
| 51 | 1.2. `outside(self, house) ∧ hold(self, key) ∧ do(go_to_house) ↝ inside(self, house)` |
| 50 | 52 | |
| 51 | 53 | ```scheme |
| 52 | 54 | (BackPredictiveImplicationScopeLink (stv 1 0.00621118) |
| … | … | @@ -85,7 +87,7 @@ The three optimal action plans required to get a reward are: |
| 85 | 87 | ) ; [a2984b519c1ddc4f][1] |
| 86 | 88 | ``` |
| 87 | 89 | |
| 88 | | 3. `inside(self, house) ∧ do(go_to_diamonds) ↝ Reward(1)` |
| 90 | 1.3. `inside(self, house) ∧ do(go_to_diamonds) ↝ Reward(1)` |
| 89 | 91 | |
| 90 | 92 | ```scheme |
| 91 | 93 | (BackPredictiveImplicationScopeLink (stv 1 0.00621118) |
| … | … | @@ -114,6 +116,134 @@ The three optimal action plans required to get a reward are: |
| 114 | 116 | ) ; [8a298365b46b204b][1] |
| 115 | 117 | ``` |
| 116 | 118 | |
| 119 | 2. Polyaction plans |
| 120 | |
| 121 | 2.1.`(outside(self, house) ∧ do(go_to_key) ≺ do(go_to_house)) ≺ do(go_to_diamonds) ↝ Reward(1)` |
| 122 | |
| 123 | ```scheme |
| 124 | (BackPredictiveImplicationScopeLink (stv 1 0.00503106) |
| 125 | (VariableSet |
| 126 | ) ; [80075fffffff523a][3] |
| 127 | (SLink |
| 128 | (ZLink |
| 129 | ) ; [800fbffffffe8ce4][3] |
| 130 | ) ; [da5f815ba9d4009f][3] |
| 131 | (BackSequentialAndLink |
| 132 | (SLink |
| 133 | (ZLink |
| 134 | ) ; [800fbffffffe8ce4][3] |
| 135 | ) ; [da5f815ba9d4009f][3] |
| 136 | (BackSequentialAndLink |
| 137 | (SLink |
| 138 | (ZLink |
| 139 | ) ; [800fbffffffe8ce4][3] |
| 140 | ) ; [da5f815ba9d4009f][3] |
| 141 | (AndLink (stv 0.18 0.0588235) |
| 142 | (EvaluationLink (stv 0.64 0.0588235) |
| 143 | (PredicateNode "outside") ; [72730412e28a734][3] |
| 144 | (ListLink |
| 145 | (ConceptNode "self") ; [40b11d11524bd751][3] |
| 146 | (ConceptNode "house") ; [63eb9919f37daa5f][3] |
| 147 | ) ; [aadca36fe9d1a468][3] |
| 148 | ) ; [ca0c329fb1ab493b][3] |
| 149 | (ExecutionLink |
| 150 | (SchemaNode "go_to_key") ; [7f46c329a5e57604][3] |
| 151 | ) ; [f8086e6fdf73cdf4][3] |
| 152 | ) ; [ddda31153cf2aa6d][3] |
| 153 | (ExecutionLink |
| 154 | (SchemaNode "go_to_house") ; [7e1737e3e117d059][3] |
| 155 | ) ; [93427319ec122fff][3] |
| 156 | ) ; [d526b02a321df7c7][3] |
| 157 | (ExecutionLink |
| 158 | (SchemaNode "go_to_diamonds") ; [7aee74cf6bad6442][3] |
| 159 | ) ; [a2630f96ffbe0861][3] |
| 160 | ) ; [b0235094e5713353][3] |
| 161 | (EvaluationLink (stv 0.1 0.0588235) |
| 162 | (PredicateNode "Reward") ; [155bb4d713db0d51][3] |
| 163 | (NumberNode "1") ; [2cf0956d543cff8e][3] |
| 164 | ) ; [d3cee8bdda06ffcb][3] |
| 165 | ) ; [bd6c515d92fe0be1][3] |
| 166 | ``` |
| 167 | |
| 168 | 2.2. `outside(self, house) ∧ do(go_to_key) ≺ do(go_to_house) ↝ inside(self, house)` |
| 169 | |
| 170 | ```scheme |
| 171 | (BackPredictiveImplicationScopeLink (stv 1 0.00559006) |
| 172 | (VariableSet |
| 173 | ) ; [80075fffffff523a][3] |
| 174 | (SLink |
| 175 | (ZLink |
| 176 | ) ; [800fbffffffe8ce4][3] |
| 177 | ) ; [da5f815ba9d4009f][3] |
| 178 | (BackSequentialAndLink |
| 179 | (SLink |
| 180 | (ZLink |
| 181 | ) ; [800fbffffffe8ce4][3] |
| 182 | ) ; [da5f815ba9d4009f][3] |
| 183 | (AndLink (stv 0.18 0.0588235) |
| 184 | (EvaluationLink (stv 0.64 0.0588235) |
| 185 | (PredicateNode "outside") ; [72730412e28a734][3] |
| 186 | (ListLink |
| 187 | (ConceptNode "self") ; [40b11d11524bd751][3] |
| 188 | (ConceptNode "house") ; [63eb9919f37daa5f][3] |
| 189 | ) ; [aadca36fe9d1a468][3] |
| 190 | ) ; [ca0c329fb1ab493b][3] |
| 191 | (ExecutionLink |
| 192 | (SchemaNode "go_to_key") ; [7f46c329a5e57604][3] |
| 193 | ) ; [f8086e6fdf73cdf4][3] |
| 194 | ) ; [ddda31153cf2aa6d][3] |
| 195 | (ExecutionLink |
| 196 | (SchemaNode "go_to_house") ; [7e1737e3e117d059][3] |
| 197 | ) ; [93427319ec122fff][3] |
| 198 | ) ; [d526b02a321df7c7][3] |
| 199 | (EvaluationLink (stv 0.34 0.0588235) |
| 200 | (PredicateNode "inside") ; [63398dcfcf85c8a3][3] |
| 201 | (ListLink |
| 202 | (ConceptNode "self") ; [40b11d11524bd751][3] |
| 203 | (ConceptNode "house") ; [63eb9919f37daa5f][3] |
| 204 | ) ; [aadca36fe9d1a468][3] |
| 205 | ) ; [871c182b52e89756][3] |
| 206 | ) ; [c4a99a11fb34aeb2][3] |
| 207 | ``` |
| 208 | |
| 209 | 2.3 `hold(self, key) ∧ do(go_to_house) ≺ do(go_to_diamonds) ↝ Reward(1)` |
| 210 | |
| 211 | ```scheme |
| 212 | (BackPredictiveImplicationScopeLink (stv 1 0.00559006) |
| 213 | (VariableSet |
| 214 | ) ; [80075fffffff523a][3] |
| 215 | (SLink |
| 216 | (ZLink |
| 217 | ) ; [800fbffffffe8ce4][3] |
| 218 | ) ; [da5f815ba9d4009f][3] |
| 219 | (BackSequentialAndLink |
| 220 | (SLink |
| 221 | (ZLink |
| 222 | ) ; [800fbffffffe8ce4][3] |
| 223 | ) ; [da5f815ba9d4009f][3] |
| 224 | (AndLink (stv 0.1 0.0588235) |
| 225 | (ExecutionLink |
| 226 | (SchemaNode "go_to_house") ; [7e1737e3e117d059][3] |
| 227 | ) ; [93427319ec122fff][3] |
| 228 | (EvaluationLink (stv 0.32 0.0588235) |
| 229 | (PredicateNode "hold") ; [4b1b7a8b0a4d2853][3] |
| 230 | (ListLink |
| 231 | (ConceptNode "self") ; [40b11d11524bd751][3] |
| 232 | (ConceptNode "key") ; [4d0844146f96d3][3] |
| 233 | ) ; [e7a9c95ae7484b28][3] |
| 234 | ) ; [e6a8f21e6b37d8f0][3] |
| 235 | ) ; [876645d6528a6fbb][3] |
| 236 | (ExecutionLink |
| 237 | (SchemaNode "go_to_diamonds") ; [7aee74cf6bad6442][3] |
| 238 | ) ; [a2630f96ffbe0861][3] |
| 239 | ) ; [c6e23df43eb8fbe6][3] |
| 240 | (EvaluationLink (stv 0.1 0.0588235) |
| 241 | (PredicateNode "Reward") ; [155bb4d713db0d51][3] |
| 242 | (NumberNode "1") ; [2cf0956d543cff8e][3] |
| 243 | ) ; [d3cee8bdda06ffcb][3] |
| 244 | ) ; [f4dde218e5acedc6][3] |
| 245 | ``` |
| 246 | |
| 117 | 247 | ## Usage |
| 118 | 248 | |
| 119 | 249 | 1. Clone malmo https://github.com/Microsoft/malmo |