English
Related papers

Related papers: Position Interpolation Improves ALiBi Extrapolatio…

200 papers

In Transformer-based architectures, the attention mechanism is inherently permutation-invariant with respect to the input sequence's tokens. To impose sequential order, token positions are typically encoded using a scheme with either fixed…

Machine Learning · Computer Science 2023-10-31 Giorgio Angelotti

Length extrapolation has attracted considerable attention recently since it allows transformers to be tested on longer sequences than those used in training. Previous research has shown that this property can be attained by using carefully…

Computation and Language · Computer Science 2023-12-20 Zhen Qin , Yiran Zhong , Hui Deng

Positional encoding plays a crucial role in transformers, significantly impacting model performance and length generalization. Prior research has introduced absolute positional encoding (APE) and relative positional encoding (RPE) to…

Computation and Language · Computer Science 2024-11-06 Chuanyang Zheng , Yihang Gao , Han Shi , Minbin Huang , Jingyao Li , Jing Xiong , Xiaozhe Ren , Michael Ng , Xin Jiang , Zhenguo Li , Yu Li

The last two decades have seen major developments in interpolatory methods for model reduction of large-scale linear dynamical systems. Advances of note include the ability to produce (locally) optimal reduced models at modest cost; refined…

Numerical Analysis · Mathematics 2014-09-18 Christopher Beattie , Serkan Gugercin

Position encoding recently has shown effective in the transformer architecture. It enables valuable supervision for dependency modeling between elements at different positions of the sequence. In this paper, we first investigate various…

Computation and Language · Computer Science 2023-11-09 Jianlin Su , Yu Lu , Shengfeng Pan , Ahmed Murtadha , Bo Wen , Yunfeng Liu

We predict future video frames from complex dynamic scenes, using an invertible neural network as the encoder of a nonlinear dynamic system with latent linear state evolution. Our invertible linear embedding (ILE) demonstrates successful…

Computer Vision and Pattern Recognition · Computer Science 2019-03-04 Robert Pottorff , Jared Nielsen , David Wingate

The ability to extrapolate, i.e., to make predictions on sequences that are longer than those presented as training examples, is a challenging problem for current deep learning models. Recent work shows that this limitation persists in…

Computation and Language · Computer Science 2022-11-29 Mirelle Bueno , Carlos Gemmell , Jeffrey Dalton , Roberto Lotufo , Rodrigo Nogueira

Transformer models, which leverage architectural improvements like self-attention, perform remarkably well on Natural Language Processing (NLP) tasks. The self-attention mechanism is position agnostic. In order to capture positional…

Computation and Language · Computer Science 2021-09-28 Zhiheng Huang , Davis Liang , Peng Xu , Bing Xiang

Rotary Positional Embeddings (RoPE) have become the standard for Large Language Models (LLMs) due to their ability to encode relative positions through geometric rotation. However, we identify a significant limitation we term ''Spectral…

Computation and Language · Computer Science 2026-02-02 Kanishk Awadhiya

Without positional information, attention-based Transformer neural networks are permutation-invariant. Absolute or relative positional embeddings are the most popular ways to feed Transformer models with positional information. Absolute…

Machine Learning · Computer Science 2021-11-10 Tatiana Likhomanenko , Qiantong Xu , Gabriel Synnaeve , Ronan Collobert , Alex Rogozhnikov

Image interpolation is a special case of image super-resolution, where the low-resolution image is directly down-sampled from its high-resolution counterpart without blurring and noise. Therefore, assumptions adopted in super-resolution…

Image and Video Processing · Electrical Eng. & Systems 2020-10-28 Junchao Zhang

Adaptive rational interpolation has been designed in the context of image processing as a new nonlinear technique that avoids the Gibbs phenomenon when we approximate a discontinuous function. In this work, we present a generalization to…

Numerical Analysis · Mathematics 2021-12-21 Francesc Arandiga , Dionisio F. Yanez

Rotary Position Embedding (RoPE)-extension refers to modifying or generalizing the Rotary Position Embedding scheme to handle longer sequences than those encountered during pre-training. However, current extension strategies are highly…

Computation and Language · Computer Science 2026-02-02 Qingyuan Tian , Wenhong Zhu , Xiaoran Liu , Xiaofeng Wang , Rui Wang

Multilinear interpolation is a powerful tool used in obtaining strong type boundedness for a variety of operators assuming only a finite set of restricted weak-type estimates. A typical situation occurs when one knows that a multilinear…

Functional Analysis · Mathematics 2007-05-23 Loukas Grafakos , Terence Tao

Position information is essential for language modeling. In softmax transformers, Rotary Position Embeddings (\textit{RoPE}) encode positions through \textit{fixed-angle} rotations, while in linear transformers, order is handled via…

Computation and Language · Computer Science 2026-04-27 Sajad Movahedi , Timur Carstensen , Arshia Afzal , Frank Hutter , Antonio Orvieto , Volkan Cevher

Deep sequence models typically degrade in accuracy when test sequences significantly exceed their training lengths, yet many critical tasks--such as algorithmic reasoning, multi-step arithmetic, and compositional generalization--require…

Machine Learning · Computer Science 2025-12-24 Philip Heejun Lee

We bring an abstract model theory perspective to interpolation. We ask, what is the role of interpolation in the study of extensions of first order logic, such as infinitary logics, generalized quantifiers and higher order logics? The…

Logic · Mathematics 2025-07-28 Jouko Väänänen

Rotary Position Embeddings (RoPE) have been shown to effectively encode positional information in transformer-based language models. However, these models fail to generalize past the sequence length they were trained on. We present YaRN…

Computation and Language · Computer Science 2026-02-10 Bowen Peng , Jeffrey Quesnelle , Honglu Fan , Enrico Shippole

So far, expensive finetuning beyond the pretraining sequence length has been a requirement for effectively extending the context of language models (LM). In this work, we break this key bottleneck by Dropping the Positional Embeddings of…

Computation and Language · Computer Science 2025-12-16 Yoav Gelberg , Koshi Eguchi , Takuya Akiba , Edoardo Cetin

Despite the remarkable capabilities of Multimodal Large Language Models (MLLMs), they still suffer from visual fading in long-context scenarios. Specifically, the attention to visual tokens diminishes as the text sequence lengthens, leading…

Computer Vision and Pattern Recognition · Computer Science 2026-03-12 Lin Chen , Bolin Ni , Qi Yang , Zili Wang , Kun Ding , Ying Wang , Houwen Peng , Shiming Xiang