Symbolic deep learning

Trying to recreate the examples from this paper PySR is always predicting scalars as a low complexity solution, which doesn’t make much sense, can you please elaborate on that? And what is wrong why I’m unable to get the right expression?

Cycles per second: 3.050e+03
Progress: 19 / 20 total iterations (95.000%)
Hall of Fame:
-----------------------------------------
Complexity  Loss       Score     Equation
1           1.278e-01  -9.446e-02  -0.08741549
2           1.165e-01  9.256e-02  square(-0.18644808)
3           2.592e-02  1.503e+00  (x0 * -0.2923665)
5           1.682e-02  2.163e-01  ((-0.10430038 * x0) * x2)
8           1.576e-02  2.176e-02  (1.6735333 * sin((-0.067048885 * x0) * x2))

The code used to generate this is:

import numpy as np
from pysr import pysr, best

# Dataset
X = np.array(messages_over_time[-1][['dx', 'dy', 'r', 'm1', 'm2']]) # Taken from this notebook https://github.com/MilesCranmer/symbolic_deep_learning/blob/master/GN_Demo_Colab.ipynb
y = np.array(messages_over_time[-1]['e64'])

# Learn equations
equations = pysr(X, y, niterations=5,
    binary_operators=["plus", "mult" , 'sub', 'pow', 'div'],
    unary_operators=[
      "cos", "exp", "sin", 'neg', 'square', 'cube', 'exp', 
      "inv(x) = 1/x"], batching=True, batchSize=1000) 


print(best(equations))

[MilesCranmer]

I am confused about what the issue here could be. Want to email me to discuss this more? My email is miles<dot>cranmer at gmail

[abdalazizrashid]

@MilesCranmer Thanks for the response. After training for 200 epochs this what I got, what do you think? Also, all this is a reproduction of what’s in the notebook.