Spaces:
Sleeping
Sleeping
MilesCranmer
commited on
Commit
•
264634b
1
Parent(s):
a95ae71
Performance ideas
Browse files
README.md
CHANGED
@@ -147,10 +147,15 @@ pd.DataFrame, Results dataframe, giving complexity, MSE, and equations
|
|
147 |
|
148 |
# TODO
|
149 |
|
|
|
|
|
150 |
- [ ] Calculate feature importances of future mutations, by looking at correlation between residual of model, and the features.
|
151 |
- Store feature importances of future, and periodically update it.
|
|
|
|
|
152 |
- [ ] Consider adding mutation for constant<->variable
|
153 |
- [ ] Consider adding mutation to pass an operator in through a new binary operator (e.g., exp(x3)->plus(exp(x3), ...))
|
|
|
154 |
- [ ] Use NN to generate weights over all probability distribution conditional on error and existing equation, and train on some randomly-generated equations
|
155 |
- [ ] Performance:
|
156 |
- [ ] Use an enum for functions instead of storing them?
|
|
|
147 |
|
148 |
# TODO
|
149 |
|
150 |
+
- [ ] Record very best individual in each population, and return at end.
|
151 |
+
- [ ] Update hall of fame every iteration
|
152 |
- [ ] Calculate feature importances of future mutations, by looking at correlation between residual of model, and the features.
|
153 |
- Store feature importances of future, and periodically update it.
|
154 |
+
- [ ] Implement more parts of the original Eureqa algorithms: https://www.creativemachineslab.com/eureqa.html
|
155 |
+
- [ ] Sympy printing
|
156 |
- [ ] Consider adding mutation for constant<->variable
|
157 |
- [ ] Consider adding mutation to pass an operator in through a new binary operator (e.g., exp(x3)->plus(exp(x3), ...))
|
158 |
+
- [ ] Hierarchical model, so can re-use functional forms. Output of one equation goes into second equation?
|
159 |
- [ ] Use NN to generate weights over all probability distribution conditional on error and existing equation, and train on some randomly-generated equations
|
160 |
- [ ] Performance:
|
161 |
- [ ] Use an enum for functions instead of storing them?
|