English
Related papers

Related papers: Mixture of Chapters: Scaling Learnt Memory in Tran…

200 papers

Self-attention has recently been adopted for a wide range of sequence modeling problems. Despite its effectiveness, self-attention suffers from quadratic compute and memory requirements with respect to sequence length. Successful approaches…

Machine Learning · Computer Science 2020-10-27 Aurko Roy , Mohammad Saffar , Ashish Vaswani , David Grangier

We study the capabilities of the transformer architecture with varying depth. Specifically, we designed a novel set of sequence learning tasks to systematically evaluate and comprehend how the depth of transformer affects its ability to…

Machine Learning · Computer Science 2024-04-03 Xingwu Chen , Difan Zou

In contrast to RNNs, which compress their history into a single hidden state, Transformers can attend to all past tokens directly. However, standard Transformers rely solely on the hidden state from the previous layer to represent the…

Machine Learning · Computer Science 2025-05-29 Gleb Gerasimov , Yaroslav Aksenov , Nikita Balagansky , Viacheslav Sinii , Daniil Gavrilov

Recurrent LLM architectures have emerged as a promising approach for improving reasoning, as they enable multi-step computation in the embedding space without generating intermediate tokens. Models such as Ouro perform reasoning by…

Sparse mixture-of-experts (MoE) layers have been shown to substantially increase model capacity without a proportional increase in computational cost and are widely used in transformer architectures, where they typically replace…

Computer Vision and Pattern Recognition · Computer Science 2026-04-16 Svetlana Pavlitska , Haixi Fan , Konstantin Ditschuneit , J. Marius Zöllner

Transformer networks, driven by self-attention, are central to Large Language Models. In generative Transformers, self-attention uses cache memory to store token projections, avoiding recomputation at each time step. However, GPU-stored…

Neural and Evolutionary Computing · Computer Science 2024-11-26 Nathan Leroux , Paul-Philipp Manea , Chirag Sudarshan , Jan Finkbeiner , Sebastian Siegel , John Paul Strachan , Emre Neftci

Sparse Mixture of Experts (SMoE) enables efficient training of large language models by routing input tokens to a select number of experts. However, training SMoE remains challenging due to the issue of representation collapse. Recent…

Computation and Language · Computer Science 2025-04-01 Giang Do , Hung Le , Truyen Tran

Continual incorporation of new knowledge is essential for the long-term evolution of large language models (LLMs). Existing approaches typically rely on parameter-update algorithms to mitigate catastrophic forgetting, yet they suffer from…

Machine Learning · Computer Science 2026-05-07 Kaustubh Pethkar , Ziyang Xiong , Zuofeng Shang , Yingcong Li

Sparsely-activated Mixture-of-experts (MoE) models allow the number of parameters to greatly increase while keeping the amount of computation for a given token or a given sample unchanged. However, a poor expert routing strategy (e.g. one…

Machine Learning · Computer Science 2022-10-17 Yanqi Zhou , Tao Lei , Hanxiao Liu , Nan Du , Yanping Huang , Vincent Zhao , Andrew Dai , Zhifeng Chen , Quoc Le , James Laudon

We study the problem of efficient generative inference for Transformer models, in one of its most challenging settings: large deep models, with tight latency targets and long sequence lengths. Better understanding of the engineering…

Merging parameter-efficient task experts has recently gained growing attention as a way to build modular architectures that can be rapidly adapted on the fly for specific downstream tasks, without requiring additional fine-tuning.…

Algorithmic reasoning requires capabilities which are most naturally understood through recurrent models of computation, like the Turing machine. However, Transformer models, while lacking recurrence, are able to perform such reasoning…

Machine Learning · Computer Science 2023-05-03 Bingbin Liu , Jordan T. Ash , Surbhi Goel , Akshay Krishnamurthy , Cyril Zhang

Recent advances in large language models highlighted the excessive quadratic cost of self-attention. Despite the significant research efforts, subquadratic attention methods still suffer from inferior performance in practice. We hypothesize…

Machine Learning · Computer Science 2025-05-02 Piotr Piękos , Róbert Csordás , Jürgen Schmidhuber

Mixture-of-Experts large language models (MoE-LLMs) marks a significant step forward of language models, however, they encounter two critical challenges in practice: 1) expert parameters lead to considerable memory consumption and loading…

Machine Learning · Computer Science 2025-02-25 Wei Huang , Yue Liao , Jianhui Liu , Ruifei He , Haoru Tan , Shiming Zhang , Hongsheng Li , Si Liu , Xiaojuan Qi

Standard transformer architectures apply the same number of layers to every token regardless of contextual difficulty. We present Token-Selective Attention (TSA), a learned per-token gate on residual updates between consecutive transformer…

Machine Learning · Computer Science 2026-05-08 Ahmed Abdelmuniem Abdalla Mohammed

Large Language Models (LLMs) are typically trained on data mixtures: most data come from web scrapes, while a small portion is curated from high-quality sources with dense domain-specific knowledge. In this paper, we show that when training…

Machine Learning · Computer Science 2026-05-12 Xinran Gu , Kaifeng Lyu , Jiazheng Li , Jingzhao Zhang

This paper aims to apply two major scaling transformations from the computing packaging industry to internet routers: the heterogeneous integration of high-bandwidth memories (HBMs) and chiplets, as well as in-package optics. We propose a…

Networking and Internet Architecture · Computer Science 2026-02-12 Isaac Keslassy , Ilay Yavlovich , Jose Yallouz , Tzu-Chien Hsueh , Yeshaiahu Fainman , Bill Lin

Transformer-based models show their effectiveness across multiple domains and tasks. The self-attention allows to combine information from all sequence elements into context-aware representations. However, global and local information has…

Computation and Language · Computer Science 2022-12-09 Aydar Bulatov , Yuri Kuratov , Mikhail S. Burtsev

In classification problems with large output spaces (up to millions of labels), the last layer can require an enormous amount of memory. Using sparse connectivity would drastically reduce the memory requirements, but as we show below, it…

Machine Learning · Computer Science 2023-11-08 Erik Schultheis , Rohit Babbar

Despite their remarkable achievement, gigantic transformers encounter significant drawbacks, including exorbitant computational and memory footprints during training, as well as severe collapse evidenced by a high degree of parameter…

Machine Learning · Computer Science 2023-03-06 Tianlong Chen , Zhenyu Zhang , Ajay Jaiswal , Shiwei Liu , Zhangyang Wang