This work proposes UGE, a two-stage training strategy that progressively and stably aligns images, text, and spatial structures by combining instruction-guided contrastive learning with graph-based spatial encoding, and introduces \dataset, a spatially grounded dataset that anchors street-view images to structured spat...
Jie Zhang, Xing-Tong Yu, Yuan Fang et al.· Proceedings of the 32nd ACM...· 0 citations
This work introduces a Query-Key router that represents the expertise required by each token under the current market context as a low-dimensional query and matches it with learnable expert keys and proposes TradingMoE, a trading-oriented sparse MoE that augments a frozen dense LLM with lightweight residual experts.
Chang Zhou, Xing-Tong Yu, Minbin Huang et al.· 0 citations
Learning transferable multimodal embeddings for urban environments is challenging because urban understanding is inherently spatial, yet existing datasets and benchmarks lack explicit alignment between street-view images and urban structure. We introduce \dataset, a spatially grounded dataset that anchors street-view i...
Jie Zhang, Xingtong Yu, Yuan Fang et al.· Proceedings of the 32nd ACM...· 0 citations
Molecular graph representation learning is widely used in chemical and biomedical research, and reusing widely available and well-validated pre-trained 2D encoders, while incorporating molecular domain knowledge during downstream adaptation, offers a more practical alternative.
