Linghe Kong

This paper proposes SWRouter, a Similarity-Contractive Window Router for multi-turn large language model routing that combines a similarity-based context segmentation mechanism for prompt construction with a dual-metric evaluation framework that decouples construction accuracy from router performance.

ET-Prune achieves a remarkable balance between efficiency and accuracy, outperforming traditional pruning methods by retaining critical evidence while cutting down on unnecessary tokens.

Error accumulation in long video generation can be mitigated with a novel frequency-domain approach, achieving 24x extrapolation without the need for extensive training.

GeoFlow reveals that integrating geospatial attributes can dramatically enhance the accuracy and diversity of urban mobility predictions.

BETA achieves state-of-the-art continual learning performance with just 0.05 million trainable parameters, outperforming traditional methods by 180–3000 times in parameter efficiency.

One-step diffusion for video face restoration is now viable, achieving state-of-the-art results by learning separate spatial and temporal priors.