English
Related papers

Related papers: SSA: Improving Performance With a Better Scoring F…

200 papers

Transformers are state-of-the-art models for a variety of sequence modeling tasks. At their core is an attention function which models pairwise interactions between the inputs at every timestep. While attention is powerful, it does not…

Computation and Language · Computer Science 2021-03-23 Hao Peng , Nikolaos Pappas , Dani Yogatama , Roy Schwartz , Noah A. Smith , Lingpeng Kong

As the core operator of Transformers, Softmax Attention exhibits excellent global modeling capabilities. However, its quadratic complexity limits its applicability to vision tasks. In contrast, Linear Attention shares a similar formulation…

Computer Vision and Pattern Recognition · Computer Science 2025-08-05 Qihang Fan , Huaibo Huang , Yuang Ai , Ran He

While Transformer architecture excel at modeling long-range dependencies contributing to its widespread adoption in vision tasks the quadratic complexity of softmax-based attention mechanisms imposes a major bottleneck, particularly when…

Computer Vision and Pattern Recognition · Computer Science 2026-03-10 Yuan Cao , Dong Wang

Transformers have transformed the field of natural language processing. This performance is largely attributed to the use of stacked self-attention layers, each of which consists of matrix multiplies as well as softmax operations. As a…

Hardware Architecture · Computer Science 2021-03-18 Jacob R. Stevens , Rangharajan Venkatesan , Steve Dai , Brucek Khailany , Anand Raghunathan

The attention mechanism is a core component of the Transformer architecture. Various methods have been developed to compute attention scores, including multi-head attention (MHA), multi-query attention, group-query attention and so on. We…

The remarkable ability of transformers to learn new concepts solely by reading examples within the input prompt, termed in-context learning (ICL), is a crucial aspect of intelligent behavior. Here, we focus on understanding the learning…

Machine Learning · Computer Science 2025-10-14 Sara Dragutinović , Andrew M. Saxe , Aaditya K. Singh

Softmax Self-Attention (SSA) is a key component of Transformer architectures. However, when utilised within skipless architectures, which aim to improve representation learning, recent work has highlighted the inherent instability of SSA…

Machine Learning · Computer Science 2026-02-06 Leo Zhang , James Martens

Transformer networks have emerged as the state-of-the-art approach for natural language processing tasks and are gaining popularity in other domains such as computer vision and audio processing. However, the efficient hardware acceleration…

Hardware Architecture · Computer Science 2024-07-29 Gamze İslamoğlu , Moritz Scherer , Gianna Paulin , Tim Fischer , Victor J. B. Jung , Angelo Garofalo , Luca Benini

The core component of attention is the scoring function, which transforms the inputs into low-dimensional queries and keys and takes the dot product of each pair. While the low-dimensional projection improves efficiency, it causes…

Machine Learning · Computer Science 2025-09-10 Yilun Kuang , Noah Amsel , Sanae Lotfi , Shikai Qiu , Andres Potapczynski , Andrew Gordon Wilson

We consider class incremental learning (CIL) problem, in which a learning agent continuously learns new classes from incrementally arriving training data batches and aims to predict well on all the classes learned so far. The main challenge…

Computer Vision and Pattern Recognition · Computer Science 2022-06-22 Hongjoon Ahn , Jihwan Kwak , Subin Lim , Hyeonsu Bang , Hyojun Kim , Taesup Moon

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

Transformer attention is typically implemented using softmax normalization, which enforces attention weights with unit sum normalization. While effective in many settings, this constraint can limit flexibility in controlling attention…

Computation and Language · Computer Science 2026-02-27 Jeongin Bae , Baeseong Park , Gunho Park , Minsub Kim , Joonhyung Lee , Junhee Yoo , Sunghyeon Woo , Jiwon Ryu , Se Jung Kwon , Dongsoo Lee

In-context reinforcement learning (ICRL) studies agents that, after pretraining, adapt to new tasks by conditioning on additional context without parameter updates. Existing theoretical analyses of ICRL largely rely on linear attention,…

Machine Learning · Computer Science 2026-05-19 Zixuan Xie , Xinyu Liu , Claire Chen , Shuze Daniel Liu , Rohan Chandra , Shangtong Zhang

Modern autoregressive models rely on attention, yet the Softmax full attention in Transformers scales quadratically with sequence length. Sliding Window Attention (SWA) achieves linear-time encoding/decoding by constraining the attention…

Machine Learning · Computer Science 2026-01-08 Jiaxu Liu , Yuhe Bai , Xiangyu Yin , Christos-Savvas Bouganis

Normalization methods improve both optimization and generalization of ConvNets. To further boost performance, the recently-proposed switchable normalization (SN) provides a new perspective for deep learning: it learns to select different…

Computer Vision and Pattern Recognition · Computer Science 2019-03-12 Wenqi Shao , Tianjian Meng , Jingyu Li , Ruimao Zhang , Yudian Li , Xiaogang Wang , Ping Luo

Softmax self-attention often assigns disproportionate weight to semantically uninformative tokens such as special tokens and punctuation, a phenomenon known as attention noise. While recent methods like Cog Attention and the Differential…

Computation and Language · Computer Science 2025-08-27 Ivan Kobyzev , Abbas Ghaddar , Dingtao Hu , Boxing Chen

The Transformer architecture has become the foundation of modern deep learning, yet its core self-attention mechanism suffers from quadratic computational complexity and lacks grounding in biological neural computation. We propose Selective…

Machine Learning · Computer Science 2026-02-17 Hasi Hays

Gating mechanisms have been widely utilized, from early models like LSTMs and Highway Networks to recent state space models, linear attention, and also softmax attention. Yet, existing literature rarely examines the specific effects of…

Computation and Language · Computer Science 2025-05-13 Zihan Qiu , Zekun Wang , Bo Zheng , Zeyu Huang , Kaiyue Wen , Songlin Yang , Rui Men , Le Yu , Fei Huang , Suozhi Huang , Dayiheng Liu , Jingren Zhou , Junyang Lin

The Softmax attention mechanism in Transformer models is notoriously computationally expensive, particularly due to its quadratic complexity, posing significant challenges in vision applications. In contrast, linear attention provides a far…

Computer Vision and Pattern Recognition · Computer Science 2025-03-12 Qihang Fan , Huaibo Huang , Ran He

Softmax function is widely used in artificial neural networks for multiclass classification, multilabel classification, attention mechanisms, etc. However, its efficacy is often questioned in literature. The log-softmax loss has been shown…

Machine Learning · Computer Science 2020-11-24 Kunal Banerjee , Vishak Prasad C , Rishi Raj Gupta , Karthik Vyas , Anushree H , Biswajit Mishra