English
Related papers

Related papers: Positional Description Matters for Transformers Ar…

200 papers

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

In Transformer-based neural machine translation (NMT), the positional encoding mechanism helps the self-attention networks to learn the source representation with order dependency, which makes the Transformer-based NMT achieve…

Computation and Language · Computer Science 2020-04-09 Kehai Chen , Rui Wang , Masao Utiyama , Eiichiro Sumita

Length generalization, the ability to generalize from small training context sizes to larger ones, is a critical challenge in the development of Transformer-based language models. Positional encoding (PE) has been identified as a major…

Computation and Language · Computer Science 2023-11-08 Amirhossein Kazemnejad , Inkit Padhi , Karthikeyan Natesan Ramamurthy , Payel Das , Siva Reddy

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

Word order, an essential property of natural languages, is injected in Transformer-based neural language models using position encoding. However, recent experiments have shown that explicit position encoding is not always useful, since some…

Computation and Language · Computer Science 2022-11-09 Karim Lasri , Alessandro Lenci , Thierry Poibeau

Within numerical reasoning, understanding numbers themselves is still a challenge for existing language models. Simple generalisations, such as solving 100+200 instead of 1+2, can substantially affect model performance (Sivakumar and…

Computation and Language · Computer Science 2024-12-12 Jasivan Alex Sivakumar , Nafise Sadat Moosavi

Positional encoding (PE) underpins how permutation-invariant Transformers represent sequence order, yet how positional information is processed and stored remains poorly understood. Modern PE methods such as RoPE still struggle on tasks…

Computation and Language · Computer Science 2026-05-29 Pierre-Antoine Lequeu , Camille Barboule , Benjamin Piwowarski

In order to preserve word-order information in a non-autoregressive setting, transformer architectures tend to include positional knowledge, by (for instance) adding positional encodings to token embeddings. Several modifications have been…

Computation and Language · Computer Science 2021-09-14 Vinit Ravishankar , Anders Søgaard

Large language models often struggle with length generalization and solving complex problem instances beyond their training distribution. We present a self-improvement approach where models iteratively generate and learn from their own…

Machine Learning · Computer Science 2025-02-14 Nayoung Lee , Ziyang Cai , Avi Schwarzschild , Kangwook Lee , Dimitris Papailiopoulos

Contemporary deep learning models effectively handle languages with diverse morphology despite not being directly integrated into them. Morphology and word order are closely linked, with the latter incorporated into transformer-based models…

Computation and Language · Computer Science 2024-05-31 Poulami Ghosh , Shikhar Vashishth , Raj Dabre , Pushpak Bhattacharyya

This paper introduces a novel approach to position embeddings in transformer models, named "Exact Positional Embeddings" (ExPE). An absolute positional embedding method that can extrapolate to sequences of lengths longer than the ones it…

Computation and Language · Computer Science 2025-10-06 Aleksis Datseris , Sylvia Vassileva , Ivan Koychev , Svetla Boytcheva

Transformers often struggle with length generalization, meaning they fail to generalize to sequences longer than those encountered during training. While arithmetic tasks are commonly used to study length generalization, certain tasks are…

Machine Learning · Computer Science 2025-04-18 Hanseul Cho , Jaeyoung Cha , Srinadh Bhojanapalli , Chulhee Yun

Error correction code is a major part of the communication physical layer, ensuring the reliable transfer of data over noisy channels. Recently, neural decoders were shown to outperform classical decoding techniques. However, the existing…

Machine Learning · Computer Science 2022-03-30 Yoni Choukroun , Lior Wolf

Causal transformer language models (LMs), such as GPT-3, typically require some form of positional encoding, such as positional embeddings. However, we show that LMs without any explicit positional encoding are still competitive with…

Computation and Language · Computer Science 2022-12-07 Adi Haviv , Ori Ram , Ofir Press , Peter Izsak , Omer Levy

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

Large language models (LLMs) frequently make errors when handling even simple numerical problems, such as comparing two small numbers. A natural hypothesis is that these errors stem from how LLMs represent numbers, and specifically, whether…

Machine Learning · Computer Science 2025-02-04 Amit Arnold Levy , Mor Geva

Transformer-based language models often achieve strong results on mathematical reasoning benchmarks while remaining fragile on basic numerical understanding and arithmetic operations. A central limitation is that numbers are processed as…

Computation and Language · Computer Science 2026-01-15 Andreea Dutulescu , Stefan Ruseti , Mihai Dascalu

Transformers have revolutionized the world of deep learning, specially in the field of natural language processing. Recently, the Audio Spectrogram Transformer (AST) was proposed for audio classification, leading to state of the art results…

Sound · Computer Science 2023-10-09 Leonardo Pepino , Pablo Riera , Luciana Ferrer

Grokking in transformers trained on algorithmic tasks is characterized by a long delay between training-set fit and abrupt generalization, but the source of that delay remains poorly understood. In encoder-decoder arithmetic models, we…

Machine Learning · Computer Science 2026-04-16 Laura Gomezjurado Gonzalez

Transformer architectures rely on explicit position encodings in order to preserve a notion of word order. In this paper, we argue that existing work does not fully utilize position information. For example, the initial proposal of a…

Computation and Language · Computer Science 2020-09-30 Zhiheng Huang , Davis Liang , Peng Xu , Bing Xiang