Search papers, labs, and topics across Lattice.
This paper introduces History-informed Lagrangian Neural Networks (HiLNN), which enhance the forecasting of mechanical systems by inferring hidden velocities and adapting to changing parameters using a recurrent encoder that extracts latent context from temporal position sequences. HiLNN overcomes the limitations of traditional Lagrangian Neural Networks by enabling the reconstruction of unobserved initial velocities and modulating system parameters like mass and damping coefficients. Empirical results demonstrate that HiLNN significantly outperforms state-of-the-art methods in long-term prediction accuracy while preserving energy consistency across various system types.
HiLNN achieves superior long-term predictions by intelligently inferring hidden dynamics from position data, outperforming traditional models that rely on complete state inputs.
Forecasting the long-horizon evolution of mechanical systems from position-only observations is a pivotal yet difficult task, as hidden velocities and trajectory-specific physical properties must be inferred simultaneously. Although physics-guided neural networks like Lagrangian Neural Networks (LNNs) guarantee physical plausibility, they generally require complete state inputs and lack adaptability to changing system parameters. To break these limitations, we introduce History-informed Lagrangian Neural Networks (HiLNN). Grounded in the insight that temporal position sequences implicitly encode underlying dynamics, HiLNN employs a recurrent encoder to extract a latent context from history. This context not only reconstructs the unobserved initial velocity but also adaptively modulates the mass matrix, potential energy, and damping coefficients of a structured Lagrangian system. By leveraging a differentiable RK4 rollout scheme, the entire pipeline is optimized end-to-end under multi-step trajectory supervision and energy-consistency regularization. Empirical evaluations across conservative, dissipative, and heterogeneous variable-parameter systems show that HiLNN delivers superior long-term prediction accuracy and maintains precise energy profiles compared to state-of-the-art baselines. The source code is publicly available at https://github.com/yingtian22/History-informed-LNN.
