Identifying Latent Neural Dynamics with Recognition-Parametrised Gaussian Process Dynamical Systems
arXiv 第 2611.17487v1 版发布一篇新论文《Identifying Latent Neural Dynamics with Recognition-Parametrised Gaussian Process Dynamical Systems》,该论文关注神经回路这一高度互联系统,并指出许多神经计算被认为由潜在神经状态的动态演化所介导,因此需要从观测数据中辨识这些潜在动力学。论文提出以识别参数化的高斯过程动力系统对潜在神经动力学进行建模与辨识。目前该文以 arXiv 新投稿(Announce Type: new)形式公开,尚处于预印本阶段。
How this story unfolded
Key points7 items Model summary · ordered by when the evidence was published
- Point 01 11/16, 13:00
论文以 arXiv 编号 2611.17487v1 发布,类型标注为 new。
Based on 1 - Point 03 11/16, 13:00
论文指出许多神经计算被认为由潜在状态的动态演化所介导。
Based on 1 - Point 04 11/16, 13:00
论文主题为用识别参数化的高斯过程动力系统辨识潜在神经动力学。
Based on 1 - Point 05 11/16, 13:00
同期另一篇 arXiv 论文(2611.18396v1)研究 grokking 是否改变训练动力学。
Based on 1 - Point 06 11/16, 13:00
另一篇论文(2606.28228v3)讨论通过扩散偏移实现潜在 SDE 的可辨识性。
Based on 1 - Point 07 11/16, 13:00
另有论文(2611.17503v1)提出动力系统算子推断的多保真度公式。
Based on 1
ReportsBy publish time · newer further down
- November 16, 2026 · Mon
- Start
Deep Neural Networks Training Efficiency: Does Grokking Modify its Dynamics?
Original summary arXiv:2611.18396v1 Announce Type: new Abstract: We study whether grokking, the delayed emergence of out-of-distribution generalization after apparent overfitting, alters training e
arXiv cs.LG Original ↗ - Start
Identifying Latent Neural Dynamics with Recognition-Parametrised Gaussian Process Dynamical Systems
Original summary arXiv:2611.17487v1 Announce Type: new Abstract: Neural circuits are densely interconnected systems, and many neural computations are thought to be mediated by the dynamical evoluti
arXiv cs.LG Original ↗ - Start
Disentangling Continuous-Time Latent Dynamics: Identifiability of Latent SDEs via Diffusion Shifts
Original summary arXiv:2606.28228v3 Announce Type: replace-cross Abstract: Causal representation learning for time series has developed strong identifiability results in discrete-time latent causal
arXiv stat.ML Original ↗ - Start
Multifidelity Operator Inference for Dynamical Systems
Original summary arXiv:2611.17503v1 Announce Type: new Abstract: This paper presents a multifidelity (MF) formulation of Operator Inference (OpInf) for learning non-intrusive, physics-based reduced
- Still developing · later reports will keep extending this axis