TL;DR
A two-stage framework that tightly couples self-supervised graph learning with dual-projection prompting that guarantees that local context is derived from the same clustering signal used during pretraining, and global and local semantics are aligned through purpose-built pathways is proposed.
Visual Question Answering (VQA) aims to achieve cross-modal semantic understanding through joint modeling of visual content and natural language. Although existing attention-based approaches effectively align features, they struggle to simultaneously accommodate global semantic modeling and local fine-grained perceptio...
Multimodal large language models process the complete visual token sequence, increasing prefill cost even when many tokens are irrelevant to the instruction. Token-wise saliency captures limited inter-token dependence; graph-based selection instead depends on how relations govern importance propagation and ranking. We...
Text-Attributed Graphs (TAGs) are prevalent in various real-world scenarios, where each node is associated with a text attribute. Representation learning on TAGs relies on a comprehensive understanding of both the textual attributes and the topological connections. Recent works have enhanced graph neural networks (GNNs...
Lightweight sentence embedding models are critical for large-scale retrieval and semantic matching, yet aggressive compression often disrupts the semantic geometry learned by powerful teachers. Existing distillation methods primarily rely on point-wise feature matching or task-specific supervision, which fails to prese...
SCSP is a training-free framework that leverages semantic compression for informative token selection in long-context embedding and can be integrated into both zero-shot and fine-tuned models in a plug-and-play manner, consistently improving their performance.
While LLMs demonstrate strong capabilities in English Relation Extraction (RE) tasks, a significant gap remains in assessing their RE performance in low-resource languages, due to a lack of standardized benchmarks. Creating native, language-specific annotated RE benchmarks is expensive and requires human expertise, whi...
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
