English
Related papers

Related papers: Randomized Positional Encodings Boost Length Gener…

200 papers

Sequence-to-sequence transduction is the core problem in language processing applications as diverse as semantic parsing, machine translation, and instruction following. The neural network models that provide the dominant solution to these…

Computation and Language · Computer Science 2021-06-09 Ekin Akyürek , Jacob Andreas

Transformer models are permutation equivariant. To supply the order and type information of the input tokens, position and segment embeddings are usually added to the input. Recent works proposed variations of positional encodings with…

Computation and Language · Computer Science 2021-11-04 Pu-Chin Chen , Henry Tsai , Srinadh Bhojanapalli , Hyung Won Chung , Yin-Wen Chang , Chun-Sung Ferng

Positional encodings enable Transformers to incorporate sequential information, yet their theoretical understanding remains limited to two properties: distance attenuation and translation invariance. Because natural language lacks purely…

Machine Learning · Computer Science 2026-02-11 Zihan Gu , Ruoyu Chen , Han Zhang , Hua Zhang , Yue Hu

Long-context modeling presents a significant challenge for transformer-based large language models (LLMs) due to the quadratic complexity of the self-attention mechanism and issues with length extrapolation caused by pretraining exclusively…

Computation and Language · Computer Science 2024-05-24 Chenghao Yang , Zi Yang , Nan Hua

Pre-trained language models demonstrate general intelligence and common sense, but long inputs quickly become a bottleneck for memorizing information at inference time. We resurface a simple method, Memorizing Transformers (Wu et al.,…

Machine Learning · Computer Science 2024-06-05 Phoebe Klett , Thomas Ahle

Autoregressive (AR) Transformer-based sequence models are known to have difficulty generalizing to sequences longer than those seen during training. When applied to text-to-speech (TTS), these models tend to drop or repeat words or produce…

Computation and Language · Computer Science 2025-03-13 Eric Battenberg , RJ Skerry-Ryan , Daisy Stanton , Soroosh Mariooryad , Matt Shannon , Julian Salazar , David Kao

In this work, we investigate the performance of untrained randomly initialized encoders in a general class of sequence to sequence models and compare their performance with that of fully-trained encoders on the task of abstractive…

Computation and Language · Computer Science 2020-02-24 Jonathan Pilault , Jaehong Park , Christopher Pal

Transformer has shown promising results in many sequence to sequence transformation tasks recently. It utilizes a number of feed-forward self-attention layers to replace the recurrent neural networks (RNN) in attention-based encoder decoder…

Computation and Language · Computer Science 2020-12-01 Pan Zhou , Ruchao Fan , Wei Chen , Jia Jia

Transformers require positional encodings to represent sequence order, yet most prior work focuses on designing new positional encodings rather than examining how positional information is fused with token embeddings. In this paper, we…

Machine Learning · Computer Science 2026-01-12 Mohamed Amine Hallam , Kuo-Kun Tseng

Deep learning models have shown state-of-the-art performance in many inverse reconstruction problems. However, it is not well understood what properties of the latent representation may improve the generalization ability of the network.…

Machine Learning · Computer Science 2018-10-16 Sandesh Ghimire , Prashnna Kumar Gyawali , John L Sapp , Milan Horacek , Linwei Wang

Transformers struggle with length generalisation, displaying poor performance even on basic tasks. We test whether these limitations can be explained through two key failures of the self-attention mechanism. The first is the inability to…

Machine Learning · Computer Science 2025-10-07 Mattia Opper , Roland Fernandez , Paul Smolensky , Jianfeng Gao

Preventing the performance decay of Transformers on inputs longer than those used for training has been an important challenge in extending the context length of these models. Though the Transformer architecture has fundamentally no limits…

We study the problem of length generalization (LG) in transformers: the ability of a model trained on shorter sequences to maintain performance when evaluated on much longer, previously unseen inputs. Prior work by Huang et al. (2025)…

Machine Learning · Computer Science 2025-11-03 Zachary Izzo , Eshaan Nichani , Jason D. Lee

Graph Neural Networks are notorious for its memory consumption. A recent Transformer-based GNN called Graph Transformer is shown to obtain superior performances when long range dependencies exist. However, combining graph data and…

Machine Learning · Computer Science 2024-03-25 Eugene Ku

Since self-attention layers in Transformers are permutation invariant by design, positional encodings must be explicitly incorporated to enable spatial understanding. However, fixed-size lookup tables used in traditional learnable position…

Machine Learning · Computer Science 2025-06-18 Huayang Li , Yahui Liu , Hongyu Sun , Deng Cai , Leyang Cui , Wei Bi , Peilin Zhao , Taro Watanabe

Ever since their conception, Transformers have taken over traditional sequence models in many tasks, such as NLP, image classification, and video/audio processing, for their fast training and superior performance. Much of the merit is…

Machine Learning · Computer Science 2023-02-17 Hongyu Hè , Marko Kabic

Attentional mechanisms are order-invariant. Positional encoding is a crucial component to allow attention-based deep model architectures such as Transformer to address sequences or images where the position of information matters. In this…

Machine Learning · Computer Science 2021-11-10 Yang Li , Si Si , Gang Li , Cho-Jui Hsieh , Samy Bengio

Large language models have exhibited intriguing in-context learning capability, achieving promising zero- and few-shot performance without updating the parameters. However, conventional in-context learning is usually restricted by length…

Computation and Language · Computer Science 2022-12-14 Yaru Hao , Yutao Sun , Li Dong , Zhixiong Han , Yuxian Gu , Furu Wei

Unsupervised meta-learning aims to learn feature representations from unsupervised datasets that can transfer to downstream tasks with limited labeled data. In this paper, we propose a novel approach to unsupervised meta-learning that…

Machine Learning · Computer Science 2025-02-11 Anna Vettoruzzo , Lorenzo Braccaioli , Joaquin Vanschoren , Marlena Nowaczyk

Training autonomous agents that can learn new tasks from only a handful of demonstrations is a long-standing problem in machine learning. Recently, transformers have been shown to learn new language or vision tasks without any weight…

Machine Learning · Computer Science 2023-12-08 Sharath Chandra Raparthy , Eric Hambro , Robert Kirk , Mikael Henaff , Roberta Raileanu