In financial trading, factor models are widely used to price assets and capture excess returns from mispricing. Recently, we have witnessed the rise of variational autoencoder-based latent factor models, which learn latent factors self-adaptively. While these models focus on modeling overall market conditions, they often fail to effectively capture the temporal patterns of individual stocks. Additionally, representing multiple factors as single values simplifies the model but limits its ability to capture complex relationships and dependencies. As a result, the learned factors are of low quality and lack diversity, reducing their effectiveness and robustness across different trading periods. To address these issues, we propose a Spatio-Temporal factOR Model based on dual vector quantized variational autoencoders, named STORM, which extracts features of stocks from temporal and spatial perspectives, then fuses and aligns these features at the fine-grained and semantic level, and represents the factors as multi-dimensional embeddings. The discrete codebooks cluster similar factor embeddings, ensuring orthogonality and diversity, which helps distinguish between different factors and enables factor selection in financial trading. To show the performance of the proposed factor model, we apply it to two downstream experiments: portfolio management on two stock datasets and individual trading tasks on six specific stocks. The extensive experiments demonstrate STORM's flexibility in adapting to downstream tasks and superior performance over baseline models.

Autonomous valet parking enables vehicles to identify parking spaces and park without human intervention, with accurate localization being a fundamental prerequisite. Existing methods typically rely on visual feature maps or semantic maps for localization. However, visual feature maps often lack robustness in underground parking environments due to similar structures, weak textures, and fluctuating lighting conditions. Semantic maps require complex post-processing and suffer from heterogeneous data association problems. In this paper, we propose directly to regress the semantic corner points to build the semantic map. Furthermore, we introduce a novel map update and merge method, which is unaffected by environmental and temporal changes, allowing continuous update and refinement of the map. To establish a global semantic map, we use four fisheye cameras to synthesize surround-view images, combined with an IMU (Inertial Measurement Unit) and wheel encoders. Real-world experiments validate the localization accuracy of the proposed system. The experimental results demonstrate the robustness and practicability of the proposed system.