🏢 Beijing University of Posts and Telecommunications

GLAFF: A novel framework that significantly improves time series forecasting robustness by fusing global timestamp information with local observations, achieving 12.5% average performance enhancement.

Revolutionizing image exposure assessment, Pixel-level IEA Network (P-IEANet) achieves state-of-the-art performance with a novel pixel-level approach, a new dataset (IEA40K), and a benchmark of 19 met…

DEPICT: A new white-box decoder for Transformer-based semantic segmentation, achieving better performance with fewer parameters by leveraging the principle of compression and connecting Transformer de…

Physics-constrained learning significantly boosts optical neural network accuracy by addressing systematic physical errors, achieving state-of-the-art results on image classification tasks.

Lumina-Next supercharges image generation: faster, more efficient, and better resolution with new architecture and sampling techniques.

LT-Defense: a searching-free backdoor defense for language models leveraging the long-tailed effect of poisoned data. It achieves 98% accuracy across 1440 models with less than 1% time cost of existin…

LucidDrag: Semantic-aware dragging transforms image editing with an intention reasoner and collaborative guidance, achieving superior accuracy, image fidelity, and semantic diversity.

FedEgoists: A novel FL collaboration formation strategy mitigating free-riders & conflicts in cross-silo business settings, ensuring optimal coalition formation for improved model performance.

FM-Delta: Lossless compression halves cloud storage for massive fine-tuned language models, saving costs without sacrificing accuracy.

Animal-Bench, a new benchmark, comprehensively evaluates multimodal video models for animal-centric video understanding, featuring 13 diverse tasks across 7 animal categories and 819 species.

Researchers refine Pearson’s correlation coefficient to precisely measure arbitrary monotone dependence, expanding its applicability beyond linear relationships.