10 papers
MATCH: Modulating Attention via In-Context Retrieval for Long-Context Transformers
Linrui Ma, Chun Hei Lo, Xinyu Wang +12
The quadratic computational cost of traditional attention mechanisms poses a major bottleneck to the scalability and practical deployment of large language models (LLMs), particula…
Near-Policy: Accelerating On-Policy Distillation via Asynchronous Generation and Selective Packing
Miao Rang, Zhenni Bi, Hang Zhou +6
Standard knowledge distillation for autoregressive models often suffers from distribution mismatch. While on-policy methods mitigate this by leveraging student-generated outputs, t…
PPE: Positional Preservation Embedding for Token Compression in Multimodal Large Language Models
Mouxiao Huang, Borui Jiang, Dehua Zheng +3
Multimodal large language models (MLLMs) have achieved strong performance on vision-language tasks, yet often suffer from inefficiencies due to redundant visual tokens. Existing to…
From Next-Token to Next-Block: A Principled Adaptation Path for Diffusion LLMs
Yuchuan Tian, Yuchen Liang, Shuo Zhang +10
Diffusion Language Models (DLMs) enable fast generation, yet training large DLMs from scratch is costly. As a practical shortcut, adapting off-the-shelf Auto-Regressive (AR) model…
Physics-Guided Multimodal Transformers are the Necessary Foundation for the Next Generation of Meteorological Science
Jing Han, Hanting Chen, Kai Han +4
This position paper argues that the next generation of artificial intelligence in meteorological and climate sciences must transition from fragmented hybrid heuristics toward a uni…
Nexus: Higher-Order Attention Mechanisms in Transformers
Hanting Chen, Chong Zhu, Kai Han +6
Transformers have achieved significant success across various domains, relying on self-attention to capture dependencies. However, the standard first-order attention mechanism is o…