English
Related papers

Related papers: Unveiling Transformer Perception by Exploring Inpu…

200 papers

Manifold models provide low-dimensional representations that are useful for processing and analyzing data in a transformation-invariant way. In this paper, we study the problem of learning smooth pattern transformation manifolds from image…

Computer Vision and Pattern Recognition · Computer Science 2013-05-20 Elif Vural , Pascal Frossard

In this paper, we work on a sound recognition system that continually incorporates new sound classes. Our main goal is to develop a framework where the model can be updated without relying on labeled data. For this purpose, we propose…

Audio and Speech Processing · Electrical Eng. & Systems 2023-01-11 Zhepei Wang , Cem Subakan , Xilin Jiang , Junkai Wu , Efthymios Tzinis , Mirco Ravanelli , Paris Smaragdis

A method is presented for accelerating inference in transformer language models by exploiting the low effective rank of the token activation manifold at each layer. The method decomposes each activation vector into a subspace component and…

Machine Learning · Computer Science 2026-05-06 Stephen J. Thomas

Among plenty of applications, low-dimensional homogeneous spaces appear in cosmological models as both, classical factor spaces of multidimensional geometry and minisuperspaces in canonical quantization. Here a new tool to restrict their…

General Relativity and Quantum Cosmology · Physics 2016-08-31 M. Rainer

Manifold-learning techniques are routinely used in mining complex spatiotemporal data to extract useful, parsimonious data representations/parametrizations; these are, in turn, useful in nonlinear model identification tasks. We focus here…

Input space reconstruction is an attractive representation learning paradigm. Despite interpretability of the reconstruction and generation, we identify a misalignment between learning by reconstruction, and learning for perception. We show…

Computer Vision and Pattern Recognition · Computer Science 2024-02-20 Randall Balestriero , Yann LeCun

We study the emergence of multi-step reasoning in deep Transformer language models through a geometric and statistical-physics lens. Treating the hidden-state trajectory as a flow on an implicit Riemannian manifold, we analyze the layerwise…

Machine Learning · Computer Science 2026-01-29 Faruk Alpay , Bugra Kilictas

Classification is a machine learning method used in many practical applications: text mining, handwritten character recognition, face recognition, pattern classification, scene labeling, computer vision, natural langage processing. A…

Machine Learning · Computer Science 2025-11-05 Doulaye Dembélé

Language models based on the Transformer architecture achieve excellent results in many language-related tasks, such as text classification or sentiment analysis. However, despite the architecture of these models being well-defined, little…

Computation and Language · Computer Science 2025-04-14 Miguel López-Otal , Jorge Gracia , Jordi Bernad , Carlos Bobed , Lucía Pitarch-Ballesteros , Emma Anglés-Herrero

Many music theoretical constructs (such as scale types, modes, cadences, and chord types) are defined in terms of pitch intervals---relative distances between pitches. Therefore, when computer models are employed in music tasks, it can be…

Sound · Computer Science 2019-02-05 Stefan Lattner , Maarten Grachten , Gerhard Widmer

Standard fine-tuning of pre-trained audio models couples representation learning with classifier training, which can obscure the true quality of the learned representations. In this work, we advocate for a disentangled two-stage framework…

Sound · Computer Science 2025-09-23 Yang Wang , Qibin Liang , Chenghao Xiao , Yizhi Li , Noura Al Moubayed , Chenghua Lin

Deep learning models develop successive representations of their input in sequential layers, the last of which maps the final representation to the output. Here we investigate the informational content of these representations by observing…

Computer Vision and Pattern Recognition · Computer Science 2023-02-28 Benjamin L. Badger

The Transformer architecture has revolutionized the field of sequence modeling and underpins the recent breakthroughs in large language models (LLMs). However, a comprehensive mathematical theory that explains its structure and operations…

Machine Learning · Computer Science 2026-04-14 Xue-Cheng Tai , Hao Liu , Lingfeng Li , Raymond H. Chan

Several recent Transformer architectures expose later layers to representations computed in the earliest layers, motivated by the observation that low-level features can become harder to recover as the residual stream is repeatedly…

Machine Learning · Computer Science 2026-05-07 Skye Gunasekaran , Téa Wright , Rui-Jie Zhu , Jason Eshraghian

We present a one-to-one correspondence between equivalence classes of embeddings of a manifold (into a larger manifold of the same dimension) and equivalence classes of certain distances on the manifold. This correspondence allows us to use…

General Relativity and Quantum Cosmology · Physics 2011-04-19 Ben E. Whale , Susan M. Scott

This paper proposes a deep representation learning using an information-theoretic loss with an aim to increase the inter-class distances as well as within-class similarity in the embedded space. Tasks such as anomaly and out-of-distribution…

Machine Learning · Computer Science 2022-02-08 Shin Ando

We propose a novel algorithm for supervised dimensionality reduction named Manifold Partition Discriminant Analysis (MPDA). It aims to find a linear embedding space where the within-class similarity is achieved along the direction that is…

Machine Learning · Computer Science 2020-11-24 Yang Zhou , Shiliang Sun

We propose a probabilistic interpretation of transformers as unrolled inference steps assuming a probabilistic Laplacian Eigenmaps model from the ProbDR framework. Our derivation shows that at initialisation, transformers perform "linear"…

Machine Learning · Computer Science 2025-07-29 Aditya Ravuri , Neil D. Lawrence

Autoencoding is a popular method in representation learning. Conventional autoencoders employ symmetric encoding-decoding procedures and a simple Euclidean latent space to detect hidden low-dimensional structures in an unsupervised way.…

Machine Learning · Computer Science 2024-10-07 Stefan C. Schonsheck , Scott Mahan , Timo Klock , Alexander Cloninger , Rongjie Lai

Learning embedding spaces of suitable geometry is critical for representation learning. In order for learned representations to be effective and efficient, it is ideal that the geometric inductive bias aligns well with the underlying…

Machine Learning · Computer Science 2021-03-30 Shuai Zhang , Yi Tay , Wenqi Jiang , Da-cheng Juan , Ce Zhang