English
Related papers

Related papers: Inference-time sparse attention with asymmetric in…

200 papers

The Segment Anything Model (SAM) achieves strong open-vocabulary segmentation, but its ViT-based image encoders dominate inference latency and memory. Existing activation compression methods, such as token merging, reduce the token length…

Computer Vision and Pattern Recognition · Computer Science 2026-05-19 Hoai-Chau Tran , Chi H. Nguyen , Duy M. H. Nguyen , Mathias Niepert , Fan Lai , Khoa D. Doan

Sparse attention reduces the quadratic complexity of full self-attention but faces two challenges: (1) an attention gap, where applying sparse attention to full-attention-trained models causes performance degradation due to train-inference…

Computation and Language · Computer Science 2026-02-02 Zhenyi Shen , Junru Lu , Lin Gui , Jiazheng Li , Yulan He , Di Yin , Xing Sun

Transformers are considered one of the most important deep learning models since 2018, in part because it establishes state-of-the-art (SOTA) records and could potentially replace existing Deep Neural Networks (DNNs). Despite the remarkable…

Machine Learning · Computer Science 2022-08-23 Hongwu Peng , Shaoyi Huang , Shiyang Chen , Bingbing Li , Tong Geng , Ang Li , Weiwen Jiang , Wujie Wen , Jinbo Bi , Hang Liu , Caiwen Ding

Transformers are slow and memory-hungry on long sequences, since the time and memory complexity of self-attention are quadratic in sequence length. Approximate attention methods have attempted to address this problem by trading off model…

Machine Learning · Computer Science 2022-06-24 Tri Dao , Daniel Y. Fu , Stefano Ermon , Atri Rudra , Christopher Ré

Scaling Transformers to longer sequence lengths has been a major problem in the last several years, promising to improve performance in language modeling and high-resolution image understanding, as well as to unlock new applications in…

Machine Learning · Computer Science 2023-07-18 Tri Dao

Transformer-based models have emerged as a leading architecture for natural language processing, natural language generation, and image generation tasks. A fundamental element of the transformer architecture is self-attention, which allows…

Machine Learning · Computer Science 2025-07-01 Venmugil Elango

Large language models (LLMs) now support extremely long context windows, but the quadratic complexity of vanilla attention results in significantly long Time-to-First-Token (TTFT) latency. Existing approaches to address this complexity…

Computation and Language · Computer Science 2025-09-04 Qianchao Zhu , Jiangfei Duan , Chang Chen , Siran Liu , Guanyu Feng , Xin Lv , Xiao Chuanfu , Dahua Lin , Chao Yang

Inference on large language models (LLMs) can be expensive in terms of the compute and memory costs involved, especially when long sequence lengths are used. In particular, the self-attention mechanism used in LLM inference contributes…

Machine Learning · Computer Science 2024-11-11 Prajwal Singhania , Siddharth Singh , Shwai He , Soheil Feizi , Abhinav Bhatele

Benefiting from the self-attention mechanism, Transformer models have attained impressive contextual comprehension capabilities for lengthy texts. The requirements of high-throughput inference arise as the large language models (LLMs)…

Hardware Architecture · Computer Science 2024-07-16 Huizheng Wang , Jiahao Fang , Xinru Tang , Zhiheng Yue , Jinxi Li , Yubin Qin , Sihan Guan , Qize Yang , Yang Wang , Chao Li , Yang Hu , Shouyi Yin

The KV cache in self-attention has emerged as a major bottleneck in long-context and large-batch inference for LLMs. Existing approaches often treat sparsity prediction and compression as separate modules, relying on auxiliary index…

Machine Learning · Computer Science 2026-03-17 Xu Yang , Jiapeng Zhang , Dongyang Zhao , Guo Chen , Zhuo Tang

Sequence parallelism (SP) serves as a prevalent strategy to handle long sequences that exceed the memory limit of a single device. However, for linear sequence modeling methods like linear attention, existing SP approaches do not take…

Machine Learning · Computer Science 2025-05-19 Weigao Sun , Zhen Qin , Dong Li , Xuyang Shen , Yu Qiao , Yiran Zhong

Long-context understanding is crucial for many NLP applications, yet transformers struggle with efficiency due to the quadratic complexity of self-attention. Sparse attention methods alleviate this cost but often impose static, predefined…

Computation and Language · Computer Science 2025-06-16 Hanzhi Zhang , Heng Fan , Kewei Sha , Yan Huang , Yunhe Feng

As large language models scale to longer contexts, loading the growing KV cache during attention computation becomes a critical bottleneck. Previous work has shown that attention computation is dominated by a small subset of tokens. This…

Distributed, Parallel, and Cluster Computing · Computer Science 2026-05-13 Di Liu , Ruitian Wang , Chen Chen , Mingliang Gong , Yongjie Yuan , Han Zhao , Yu Feng , Quan Chen , Minyi Guo

Attention is the cornerstone of modern Large Language Models (LLMs). Yet its quadratic complexity hinders efficiency and scalability, especially for long-context processing. A promising approach is to leverage sparsity in attention.…

Computation and Language · Computer Science 2025-02-18 Yizhao Gao , Zhichen Zeng , Dayou Du , Shijie Cao , Peiyuan Zhou , Jiaxing Qi , Junjie Lai , Hayden Kwok-Hay So , Ting Cao , Fan Yang , Mao Yang

How much information do NLP tasks really need from a transformer's attention mechanism at application-time (inference)? From recent work, we know that there is sparsity in transformers and that the floating-points within its computation can…

Computation and Language · Computer Science 2021-06-03 Tianchu Ji , Shraddhan Jain , Michael Ferdman , Peter Milder , H. Andrew Schwartz , Niranjan Balasubramanian

Transformers have achieved remarkable success in sequence modeling and beyond but suffer from quadratic computational and memory complexities with respect to the length of the input sequence. Leveraging techniques include sparse and linear…

Machine Learning · Computer Science 2022-08-02 Tan Nguyen , Richard G. Baraniuk , Robert M. Kirby , Stanley J. Osher , Bao Wang

Attention is the dominant source of latency during long-context LLM inference, an increasingly popular workload with reasoning models and RAG. We propose Kascade, a training-free sparse attention method that leverages known observations…

Machine Learning · Computer Science 2025-12-19 Dhruv Deshmukh , Saurabh Goyal , Nipun Kwatra , Ramachandran Ramjee

Transformers have become the foundation of numerous state-of-the-art AI models across diverse domains, thanks to their powerful attention mechanism for modeling long-range dependencies. However, the quadratic scaling complexity of attention…

Hardware Architecture · Computer Science 2026-01-29 Zhenkun Fan , Zishen Wan , Che-Kai Liu , Ashwin Sanjay Lele , Win-San Khwa , Bo Zhang , Meng-Fan Chang , Arijit Raychowdhury

Processing long contexts has become a critical capability for modern large language models (LLMs). However, serving long-context LLMs comes with significant inference costs due to the high memory overhead of the key-value (KV) cache.…

Machine Learning · Computer Science 2025-03-04 Qihui Zhou , Peiqi Yin , Pengfei Zuo , James Cheng

Transformers have revolutionized natural language processing, but their quadratic complexity with respect to sequence length remains a fundamental bottleneck for long-range modeling. While sparse attention mechanisms like RingAttention…

Computation and Language · Computer Science 2026-03-31 Dong Liu , Yanxuan Yu