11 papers
Last But Not Least: Boundary Attention CalibratiON for Multimodal KV Cache Compression
Tianhao Chen, Yuheng Wu, Kelu Yao +3
Multimodal Large Language Models (MLLMs) achieve strong vision-language reasoning, but long visual contexts enlarge the KV cache and increase decoding latency. Existing compression…
The Latent Space: Foundation, Evolution, Mechanism, Ability, and Outlook
Xinlei Yu, Zhangquan Chen, Yongbo He +36
Latent space is rapidly emerging as a native substrate for language-based models. While modern systems are still commonly understood through explicit token-level generation, an inc…
AHPA: Adaptive Hierarchical Prior Alignment for Diffusion Transformers
Ruibin Min, Yexin Liu, Aimin Pan +5
Representation alignment has recently emerged as an effective paradigm for accelerating Diffusion Transformer training. Despite their success, existing alignment methods typically…
PhysRVG: Physics-Aware Unified Reinforcement Learning for Video Generative Models
Qiyuan Zhang, Biao Gong, Shuai Tan +7
Physical principles are fundamental to realistic visual simulation, but remain a significant oversight in transformer-based video generation. This gap highlights a critical limitat…
Asymmetric Cross-Modal Knowledge Distillation: Bridging Modalities with Weak Semantic Consistency
Riling Wei, Kelu Yao, Chuanguang Yang +3
Cross-modal Knowledge Distillation has demonstrated promising performance on paired modalities with strong semantic connections, referred to as Symmetric Cross-modal Knowledge Dist…
Falcon: A Remote Sensing Vision-Language Foundation Model (Technical Report)
Kelu Yao, Nuo Xu, Rong Yang +8
This paper introduces a holistic vision-language foundation model tailored for remote sensing, named Falcon. Falcon offers a unified, prompt-based paradigm that effectively execute…