English
Related papers

Related papers: On Isotropy Calibration of Transformers

200 papers

Several studies have explored various advantages of multilingual pre-trained models (such as multilingual BERT) in capturing shared linguistic knowledge. However, less attention has been paid to their limitations. In this paper, we…

Computation and Language · Computer Science 2022-03-18 Sara Rajaee , Mohammad Taher Pilehvar

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

Entropy modeling is a key component for high-performance image compression algorithms. Recent developments in autoregressive context modeling helped learning-based methods to surpass their classical counterparts. However, the performance of…

Image and Video Processing · Electrical Eng. & Systems 2024-02-28 A. Burakhan Koyuncu , Han Gao , Atanas Boev , Georgii Gaikov , Elena Alshina , Eckehard Steinbach

Centralized training is the standard paradigm in deep learning, enabling models to learn from a unified dataset in a single location. In such setup, isotropic feature distributions naturally arise as a mean to support well-structured and…

Machine Learning · Computer Science 2026-02-09 Chiara Lanza , Roberto Pereira , Marco Miozzo , Eduard Angelats , Paolo Dini

Pre-trained speech Transformers have facilitated great success across various speech processing tasks. However, fine-tuning these encoders for downstream tasks require sufficiently large training data to converge or to achieve…

Computation and Language · Computer Science 2022-10-25 Hao Yang , Jinming Zhao , Gholamreza Haffari , Ehsan Shareghi

A common approach for sequence tagging tasks based on contextual word representations is to train a machine learning classifier directly on these embedding vectors. This approach has two shortcomings. First, such methods consider single…

The representation degeneration problem is a phenomenon that is widely observed among self-supervised learning methods based on Transformers. In NLP, it takes the form of anisotropy, a singular property of hidden representations which makes…

Computation and Language · Computer Science 2023-06-14 Nathan Godey , Éric de la Clergerie , Benoît Sagot

The representation degeneration problem is a phenomenon that is widely observed among self-supervised learning methods based on Transformers. In NLP, it takes the form of anisotropy, a singular property of hidden representations which makes…

Computation and Language · Computer Science 2024-01-25 Nathan Godey , Éric de la Clergerie , Benoît Sagot

I investigate the scattering properties of transformation devices as the traditional impedance matching criteria are altered. This is demonstrated using simple theory and augmented by numerical simulations that investigate the role of…

Optics · Physics 2018-05-04 Paul Kinsler

To analyze high-dimensional and complex data in the real world, deep generative models, such as variational autoencoder (VAE) embed data in a low-dimensional space (latent space) and learn a probabilistic model in the latent space. However,…

Machine Learning · Computer Science 2020-09-01 Keizo Kato , Jing Zhou , Tomotake Sasaki , Akira Nakagawa

Continuous representation of words is a standard component in deep learning-based NLP models. However, representing a large vocabulary requires significant memory, which can cause problems, particularly on resource-constrained platforms.…

Computation and Language · Computer Science 2020-01-24 Siyu Liao , Jie Chen , Yanzhi Wang , Qinru Qiu , Bo Yuan

Autoregressive language models have demonstrated a remarkable ability to extract latent structure from text. The embeddings from large language models have been shown to capture aspects of the syntax and semantics of language. But what…

Machine Learning · Computer Science 2026-01-09 Liyi Zhang , Michael Y. Li , R. Thomas McCoy , Theodore R. Sumers , Jian-Qiao Zhu , Thomas L. Griffiths

Autoencoders exhibit impressive abilities to embed the data manifold into a low-dimensional latent space, making them a staple of representation learning methods. However, without explicit supervision, which is often unavailable, the…

Machine Learning · Computer Science 2023-01-12 Felix Leeb , Stefan Bauer , Michel Besserve , Bernhard Schölkopf

Pre-trained language models such as BERT have become a more common choice of natural language processing (NLP) tasks. Research in word representation shows that isotropic embeddings can significantly improve performance on downstream tasks.…

Computation and Language · Computer Science 2021-08-30 Yuxin Liang , Rui Cao , Jie Zheng , Jie Ren , Ling Gao

Deep reinforcement learning systems often suffer from unstable training dynamics due to non-stationarity, where learning objectives and data distributions evolve over time. We show that under non-stationary targets, isotropic Gaussian…

Machine Learning · Computer Science 2026-03-20 Ali Saheb Pasand , Johan Obando-Ceron , Aaron Courville , Pouya Bashivan , Pablo Samuel Castro

A recent body of work has demonstrated that Transformer embeddings can be linearly decomposed into well-defined sums of factors, that can in turn be related to specific network inputs or components. There is however still a dearth of work…

Computation and Language · Computer Science 2023-10-12 Timothee Mickus , Raúl Vázquez

Word embedding models offer continuous vector representations that can capture rich contextual semantics based on their word co-occurrence patterns. While these word vectors can provide very effective features used in many NLP tasks such as…

Computation and Language · Computer Science 2017-02-27 Cem Safak Sahin , Rajmonda S. Caceres , Brandon Oselio , William M. Campbell

Word embeddings learnt from large corpora have been adopted in various applications in natural language processing and served as the general input representations to learning systems. Recently, a series of post-processing methods have been…

Machine Learning · Computer Science 2019-10-25 Shuai Tang , Mahta Mousavi , Virginia R. de Sa

Large language models exhibit sophisticated capabilities, yet understanding how they work internally remains a central challenge. A fundamental obstacle is that training selects for behavior, not circuitry, so many weight configurations can…

Machine Learning · Computer Science 2026-02-27 Joshua S. Schiffman

Contextual adaptation in token embeddings plays a central role in determining how well language models maintain coherence and retain semantic relationships over extended text sequences. Static embeddings often impose constraints on lexical…

Computation and Language · Computer Science 2025-03-27 Koinis Vassilis , Godfrey Milbourne , Harriet Featherstone , Xanthe Peverell , Yorick Bletchley , Zachary Montford