面向随机三维人体运动预测的高斯混合潜变量流
Gaussian-Mixture Latent Flow for Stochastic 3D Human Motion Prediction
用可逆归一化流 + 数据驱动高斯混合先验建模人体运动的随机性，并通过可处理的似然计算给出不确定性估计；做运动预测/生成、需要多样且物理合理结果的工作可直接复用这一潜变量流框架与混合先验设计。
原文依据 we propose a latent flow-based model equipped with a data-driven Gaussian mixture prior
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论随机人体运动预测旨在预测未来的运动分布。尽管近期研究在准确性和多样性方面取得了强劲性能，但它们往往忽视了合理性（例如产生物理上不现实的预测）与不确定性量化，而这两者对于实际应用和下游任务至关重要。为解决这些问题，我们提出了一种基于潜变量流的模型，并为其配备数据驱动的高斯混合先验，相比传统的单模态先验，该先验能更有效地解耦多样的人体行为。该先验从训练数据的模式中导出，无需额外标注。此外，模型完全可逆的特性使得通过可处理的似然计算实现自然的不确定性量化成为可能。在 Human3.6M 和 AMASS 数据集上的实验表明，我们的方法在准确性和合理性上均达到最优性能，同时还能提供可靠的不确定性估计。
英文原文摘要 1016 字符
Stochastic human motion prediction aims to forecast future motion distributions. Although recent studies have achieved strong performance in terms of accuracy and diversity, they often overlook plausibility (e.g., resulting in physically unrealistic predictions) and uncertainty quantification, which is essential for real-world applications and downstream tasks. To address these issues, we propose a latent flow-based model equipped with a data-driven Gaussian mixture prior that more effectively disentangles diverse human behaviors than conventional single-modal priors. This prior is derived from patterns in the training data without requiring additional annotations. Furthermore, the fully invertible nature of our model enables natural uncertainty quantification through tractable likelihood computation. Experiments on the Human3.6M and AMASS datasets demonstrate that our approach achieves state-of-the-art performance in both accuracy and plausibility, while also providing reliable uncertainty estimates.