English
Related papers

Related papers: Transformer Injectivity & Geometric Robustness - A…

200 papers

Prompt injection attacks pose significant risks to language model safety, yet existing defenses are typically evaluated using classification performance. We show that high detection performance does not imply representational robustness.…

Cryptography and Security · Computer Science 2026-05-20 Becky Mashaido , Tapadhir Das

Despite their empirical success, pushing Transformer architectures to extreme depth often leads to a paradoxical failure: representations become increasingly redundant, lose rank, and ultimately collapse. Existing explanations largely…

Machine Learning · Computer Science 2026-01-16 Haoran Su , Chenyu You

Prompt-level safety probes for large language models use hidden-state representations to separate safe from unsafe prompts, but strong average detection performance does not explain the geometry of this separation. In particular, it remains…

Machine Learning · Computer Science 2026-05-21 Woo Seob Sim , Yu Rang Park

Injectivity plays an important role in generative models where it enables inference; in inverse problems and compressed sensing with generative priors it is a precursor to well posedness. We establish sharp characterizations of injectivity…

Machine Learning · Computer Science 2021-10-12 Michael Puthawala , Konik Kothari , Matti Lassas , Ivan Dokmanić , Maarten de Hoop

We study representations of data from an arbitrary metric space $\mathcal{X}$ in the space of univariate Gaussian mixtures with a transport metric (Delon and Desolneux 2020). We derive embedding guarantees for feature maps implemented by…

Machine Learning · Computer Science 2023-10-17 Anastasis Kratsios , Valentin Debarnot , Ivan Dokmanić

Transformers enable powerful content-based global routing via self-attention, but they lack an explicit local geometric prior along the sequence axis. As a result, the placement of locality-inducing modules in hybrid architectures has…

Machine Learning · Computer Science 2026-02-17 Yukun Zhang , Xueqing Zhou

Despite the demonstrated empirical efficacy of prompt tuning to adapt a pretrained language model for a new task, the theoretical underpinnings of the difference between "tuning parameters before the input" against "the tuning of model…

Machine Learning · Computer Science 2023-11-17 Yihan Wang , Jatin Chauhan , Wei Wang , Cho-Jui Hsieh

We study permutation-invariant embeddings of $d$-dimensional point sets, which are defined by sorting $D$ independent one-dimensional projections of the input. Such embeddings arise in graph deep learning where outputs should be invariant…

Machine Learning · Computer Science 2026-05-26 Nadav Dym , Matthias Wellershoff , Efstratios Tsoukanis , Daniel Levy , Radu Balan

Several recent works demonstrate that transformers can implement algorithms like gradient descent. By a careful construction of weights, these works show that multiple layers of transformers are expressive enough to simulate iterations of…

Machine Learning · Computer Science 2023-11-13 Kwangjun Ahn , Xiang Cheng , Hadi Daneshmand , Suvrit Sra

Transformers have become the dominant architecture in modern machine learning, yet the theoretical understanding of their training dynamics remains limited. This paper develops a rigorous mathematical framework for analyzing gradient-based…

Optimization and Control · Mathematics 2026-05-19 Raphaël Barboni , Maarten V. de Hoop , Takashi Furuya , Gabriel Peyré

We prove that the geodesic X-ray transform is injective on $L^2$ when the Riemannian metric is simple but the metric tensor is only finitely differentiable. The number of derivatives needed depends explicitly on dimension, and in dimension…

Analysis of PDEs · Mathematics 2023-09-25 Joonas Ilmavirta , Antti Kykkänen , Kelvin Lam

Stability and robustness are critical for deploying Transformers in safety-sensitive settings. A principled way to enforce such behavior is to constrain the model's Lipschitz constant. However, approximation-theoretic guarantees for…

Machine Learning · Computer Science 2026-02-18 Takashi Furuya , Davide Murari , Carola-Bibiane Schönlieb

We introduce an implicit representation of continuous, bijective, orientation-preserving maps between genus zero surfaces with or without boundary. The distortion of these maps can easily be minimized by optimizing the Ginzburg-Landau…

Graphics · Computer Science 2026-05-05 Etienne Corman , Yousuf Soliman , Robin Magnet , Mark Gillespie

We introduce the Graded Transformer framework, a new class of sequence models that embeds algebraic inductive biases through grading transformations on vector spaces. Extending Graded Neural Networks (GNNs), we propose two architectures:…

Machine Learning · Computer Science 2025-09-03 Tony Shaska

Recent research has demonstrated that transformers, particularly linear attention models, implicitly execute gradient-descent-like algorithms on data provided in-context during their forward inference step. However, their capability in…

Machine Learning · Computer Science 2024-10-31 Max Vladymyrov , Johannes von Oswald , Mark Sandler , Rong Ge

Transformers are widely used to extract semantic meanings from input tokens, yet they usually operate as black-box models. In this paper, we present a simple yet informative decomposition of hidden states (or embeddings) of trained…

Machine Learning · Computer Science 2024-02-06 Jiajun Song , Yiqiao Zhong

Multi-head self-attention-based Transformers have shown promise in different learning tasks. Albeit these models exhibit significant improvement in understanding short-term and long-term contexts from sequences, encoders of Transformers and…

Computation and Language · Computer Science 2023-10-24 Ayan Sengupta , Md Shad Akhtar , Tanmoy Chakraborty

Transformer-based language models create hidden representations of their inputs at every layer, but only use final-layer representations for prediction. This obscures the internal decision-making process of the model and the utility of its…

Computation and Language · Computer Science 2024-06-21 Alexander Yom Din , Taelin Karidi , Leshem Choshen , Mor Geva

Consider a finite dimensional real vector space and a finite group acting unitarily on it. We study the general problem of constructing Euclidean stable embeddings of the quotient space of orbits. Our embedding is based on subsets of sorted…

Representation Theory · Mathematics 2025-08-18 Radu Balan , Efstratios Tsoukanis

Transformers excel at in-context learning (ICL) -- learning from demonstrations without parameter updates -- but how they do so remains a mystery. Recent work suggests that Transformers may internally run Gradient Descent (GD), a…

Machine Learning · Computer Science 2024-11-19 Deqing Fu , Tian-Qi Chen , Robin Jia , Vatsal Sharan
‹ Prev 1 2 3 10 Next ›