Search papers, labs, and topics across Lattice.
This paper introduces the Causal Adversarial Subspace Clustering (CASC) framework, which enhances deep subspace clustering for multivariate spatiotemporal data by integrating a U-Net-inspired architecture with stacked FAConvLSTM layers. By incorporating a graph attention transformer-based self-expressive network and two novel learning objectives, CASC effectively captures causal dependencies and evolving temporal dynamics, addressing the limitations of existing methods that rely on static subspace structures. The key result demonstrates that CASC significantly improves the identification of latent regimes in complex spatiotemporal systems, leading to more accurate clustering outcomes that reflect underlying causal processes.
Evolving subspace structures can be captured in spatiotemporal data through a novel causal framework that outperforms traditional clustering methods.
Deep subspace clustering plays a critical role in applications involving multivariate spatiotemporal data, such as sea ice monitoring, disease spread analysis, and tracking neuro-degeneration over time. Despite recent advances, existing methods primarily rely on geometric self-expressiveness, assume static subspace structures, and often fail to capture causal dependencies, local spatial interactions, and long-range temporal dynamics inherent in complex spatiotemporal systems. To address these limitations, we propose a novel Causal Adversarial Subspace Clustering (CASC) framework for discovering evolving latent regimes in high-dimensional spatiotemporal data. CASC integrates a U-Net-inspired deep adversarial clustering architecture with stacked FAConvLSTM layers to preserve spatial and temporal structure while learning robust latent representations. A graph attention transformer-based self-expressive network is introduced to jointly model local spatial relationships, global dependencies, and long-range temporal interactions. Furthermore, we propose two new learning objectives: (1) a Causal Subspace Preservation Loss that aligns self-expression coefficients with latent causal relationships, encouraging clusters to reflect underlying causal processes rather than simple feature similarity, and (2) a Dynamic Temporal Subspace Evolution Loss that captures evolving subspace structures and temporal regime transitions in nonstationary environments. Together, these components transform deep subspace clustering from a correlation-driven paradigm into a causal-temporal regime discovery framework.
