State Representation + Architecture¶

• Encoding:
  • Regularity, e.g. RK4
  • Symmetries, e.g. equivariant, invariant
  • Conservation Laws, e.g. conservation of mass, energy, rotational invariance
• Operator Learning
• Universal Differential Equations
  • incorporating NN into traditional PDE solvers

NerFs + Siren¶

Finite Difference + CNNs¶

Finite Volume + Neural Operators¶

Euler + ResNets¶