English
Related papers

Related papers: From Attention to Atoms: Spectral Dictionary Learn…

200 papers

Transformer models typically calculate attention matrices using dot products, which have limitations when capturing nonlinear relationships between embedding vectors. We propose Neural Attention, a technique that replaces dot products with…

Machine Learning · Computer Science 2025-11-10 Andrew DiGiugno , Ausif Mahmood

The Learnable Tree Filter presents a remarkable approach to model structure-preserving relations for semantic segmentation. Nevertheless, the intrinsic geometric constraint forces it to focus on the regions with close spatial distance,…

Computer Vision and Pattern Recognition · Computer Science 2020-12-08 Lin Song , Yanwei Li , Zhengkai Jiang , Zeming Li , Xiangyu Zhang , Hongbin Sun , Jian Sun , Nanning Zheng

Transformer architecture has become ubiquitous in the natural language processing field. To interpret the Transformer-based models, their attention patterns have been extensively analyzed. However, the Transformer architecture is not only…

Computation and Language · Computer Science 2021-09-16 Goro Kobayashi , Tatsuki Kuribayashi , Sho Yokoi , Kentaro Inui

We present a formal operator-theoretic framework for analyzing Transformer-based language models using free probability theory. By modeling token embeddings and attention mechanisms as self-adjoint operators in a tracial \( W^*…

Machine Learning · Computer Science 2025-08-19 Swagatam Das

Self-modulating mechanisms introduce dynamic adaptation capabilities within language models through contextual realignment strategies that influence token embedding trajectories across extended sequences. Contextual Flux is explored as an…

Computation and Language · Computer Science 2025-08-11 Henry Evidail , Zachary Mountebank , Alistair Hathersage , Peter Stanhope , Basil Ravenscroft , Tobias Waddingham

Deep neural networks (DNN) are quickly becoming the de facto standard modeling method for many natural language generation (NLG) tasks. In order for such models to truly be useful, they must be capable of correctly generating utterances for…

Computation and Language · Computer Science 2019-11-11 Chris Kedzie , Kathleen McKeown

We introduce adaptive input representations for neural language modeling which extend the adaptive softmax of Grave et al. (2017) to input representations of variable capacity. There are several choices on how to factorize the input and…

Computation and Language · Computer Science 2019-02-26 Alexei Baevski , Michael Auli

This paper proposes a model for transforming speech features using the frequency-directional attention model for End-to-End (E2E) automatic speech recognition. The idea is based on the hypothesis that in the phoneme system of each language,…

Audio and Speech Processing · Electrical Eng. & Systems 2022-03-30 Akihiro Dobashi , Chee Siang Leow , Hiromitsu Nishizaki

Existing large language models have to run K times to generate a sequence of K tokens. In this paper, we present RecycleGPT, a generative language model with fast decoding speed by recycling pre-generated model states without running the…

Computation and Language · Computer Science 2024-05-24 Yufan Jiang , Qiaozhi He , Xiaomin Zhuang , Zhihua Wu , Kunpeng Wang , Wenlai Zhao , Guangwen Yang

Multi-head attention empowers the recent success of transformers, the state-of-the-art models that have achieved remarkable success in sequence modeling and beyond. These attention mechanisms compute the pairwise dot products between the…

Machine Learning · Computer Science 2022-06-02 Tan Nguyen , Minh Pham , Tam Nguyen , Khai Nguyen , Stanley J. Osher , Nhat Ho

Most representation learning algorithms for language and image processing are local, in that they identify features for a data point based on surrounding points. Yet in language processing, the correct meaning of a word often depends on its…

Machine Learning · Computer Science 2014-02-19 Anjan Nepal , Alexander Yates

A major limitation for the broader scope of problems solvable by transformers is the quadratic scaling of computational complexity with input size. In this study, we investigate the recurrent memory augmentation of pre-trained transformer…

Computation and Language · Computer Science 2024-02-07 Aydar Bulatov , Yuri Kuratov , Yermek Kapushev , Mikhail S. Burtsev

Natural language processing often involves computations with semantic or syntactic graphs to facilitate sophisticated reasoning based on structural relationships. While convolution kernels provide a powerful tool for comparing graph…

Computation and Language · Computer Science 2018-02-13 Sahil Garg , Greg Ver Steeg , Aram Galstyan

We compare the fast training and decoding speed of RETURNN of attention models for translation, due to fast CUDA LSTM kernels, and a fast pure TensorFlow beam search decoder. We show that a layer-wise pretraining scheme for recurrent…

Neural and Evolutionary Computing · Computer Science 2019-08-06 Albert Zeyer , Tamer Alkhouli , Hermann Ney

We introduce a new beam search decoder that is fully differentiable, making it possible to optimize at training time through the inference procedure. Our decoder allows us to combine models which operate at different granularities (e.g.…

Computation and Language · Computer Science 2019-02-19 Ronan Collobert , Awni Hannun , Gabriel Synnaeve

Knowing which words have been attended to in previous time steps while generating a translation is a rich source of information for predicting what words will be attended to in the future. We improve upon the attention model of Bahdanau et…

Neural and Evolutionary Computing · Computer Science 2016-07-19 Zichao Yang , Zhiting Hu , Yuntian Deng , Chris Dyer , Alex Smola

Transformer-based pre-trained language models such as BERT have achieved remarkable results in Semantic Sentence Matching. However, existing models still suffer from insufficient ability to capture subtle differences. Minor noise like word…

Computation and Language · Computer Science 2023-04-17 Sirui Wang , Di Liang , Jian Song , Yuntao Li , Wei Wu

Transformer architectures are the backbone of the modern AI revolution. However, they are based on simply stacking the same blocks in dozens of layers and processing information sequentially from one block to another. In this paper, we…

Computation and Language · Computer Science 2024-12-24 Prateek Verma , Mert Pilanci

Modern language models rely on static vocabularies, fixed before pretraining, in contrast to the adaptive vocabulary acquisition observed in human language learning. To bridge this gap, we introduce vocabulary curriculum learning, an…

Computation and Language · Computer Science 2025-02-26 Fangyuan Yu

We introduce a fully spectral learning framework that eliminates traditional neural layers by operating entirely in the wavelet domain. The model applies learnable nonlinear transformations, including soft-thresholding and gain-phase…

Machine Learning · Computer Science 2025-07-29 Andrew Kiruluta