The core idea of our EGODE is to understand the continuous evolution and instantaneous changes in rigid-body systems.

However, most state-of-art neural models are known to be data-hungry and require substantial training data.

Where the updated knowledge resides in memories is a fundamental question for model editing.