Physics-inspired neural networks (NNs), such as Hamiltonian or Lagrangian
NNs, dramatically outperform other learned dynamics models by leveraging strong
inductive biases. These models, however, are challenging to apply to many real
world systems, such as those that don't conserve energy or contain contacts, a
common setting for robotics and reinforcement learning. In this paper, we
examine the inductive biases that make physics-inspired models successful in
practice. We show that, contrary to conventional wisdom, the improved
generalization of HNNs is the result of modeling acceleration directly and
avoiding artificial complexity from the coordinate system, rather than
symplectic structure or energy conservation. We show that by relaxing the
inductive biases of these models, we can match or exceed performance on
energy-conserving systems while dramatically improving performance on
practical, non-conservative systems. We extend this approach to constructing
transition models for common Mujoco environments, showing that our model can
appropriately balance inductive biases with the flexibility required for
model-based control.

研究物理启发的神经网络的诱导偏差及其应用。表明与常规认识相反，通过直接建模加速度避免人工坐标系的人工复杂性，而不是辛结构或能量守恒，改善了 HNN 的广义性能。在实际中，通过放松这些模型的诱导偏差，可以在能量守恒系统上匹配或超过其性能，同时显著提高非守恒系统上的性能。作者将这种方法扩展到通用 Mujoco 环境的转换模型构建中，展示了我们的模型可以适当地平衡诱导偏差与需求灵活性，从而实现基于模型的控制。