English
Related papers

Related papers: From Shortcut to Induction Head: How Data Diversit…

200 papers

Deep learning models often rely only on a small set of features even when there is a rich set of predictive signals in the training data. This makes models brittle and sensitive to distribution shifts. In this work, we first examine vision…

Computer Vision and Pattern Recognition · Computer Science 2023-09-01 Armand Mihai Nicolicioiu , Andrei Liviu Nicolicioiu , Bogdan Alexe , Damien Teney

Large transformer-based models are able to perform in-context few-shot learning, without being explicitly trained for it. This observation raises the question: what aspects of the training regime lead to this emergent behavior? Here, we…

We study how multi-head softmax attention models are trained to perform in-context learning on linear data. Through extensive empirical experiments and rigorous theoretical analysis, we demystify the emergence of elegant attention patterns:…

Machine Learning · Computer Science 2025-05-29 Jianliang He , Xintian Pan , Siyu Chen , Zhuoran Yang

Deep-learning models can extract a rich assortment of features from data. Which features a model uses depends not only on \emph{predictivity} -- how reliably a feature indicates training-set labels -- but also on \emph{availability} -- how…

Machine Learning · Computer Science 2024-07-15 Katherine L. Hermann , Hossein Mobahi , Thomas Fel , Michael C. Mozer

Transformers have theoretical limitations in modeling certain sequence-to-sequence tasks, yet it remains largely unclear if these limitations play a role in large-scale pretrained LLMs, or whether LLMs might effectively overcome these…

Machine Learning · Computer Science 2025-10-24 Mayank Jobanputra , Yana Veitsman , Yash Sarrof , Aleksandra Bakalova , Vera Demberg , Ellie Pavlick , Michael Hahn

Graph Transformers, which incorporate self-attention and positional encoding, have recently emerged as a powerful architecture for various graph learning tasks. Despite their impressive performance, the complex non-convex interactions…

Machine Learning · Computer Science 2024-06-05 Hongkang Li , Meng Wang , Tengfei Ma , Sijia Liu , Zaixi Zhang , Pin-Yu Chen

Transformer has become the dominant architecture for sequence modeling, yet a detailed understanding of how its structural parameters influence expressive power remains limited. In this work, we study the approximation properties of…

Machine Learning · Computer Science 2026-04-01 Penghao Yu , Haotian Jiang , Zeyu Bao , Ruoxi Yu , Qianxiao Li

The capacity of neural networks like the widely adopted transformer is known to be very high. Evidence is emerging that they learn successfully due to inductive bias in the training routine, typically a variant of gradient descent (GD). To…

Machine Learning · Computer Science 2023-03-09 William Merrill , Vivek Ramanujan , Yoav Goldberg , Roy Schwartz , Noah Smith

Transformers use the dense self-attention mechanism which gives a lot of flexibility for long-range connectivity. Over multiple layers of a deep transformer, the number of possible connectivity patterns increases exponentially. However,…

Machine Learning · Computer Science 2023-06-05 Md Shamim Hussain , Mohammed J. Zaki , Dharmashankar Subramanian

This paper investigates the learning theory of Transformer networks for regression tasks on the compact Euclidean domain $[0,1]^d$ and $d$-dimensional compact Riemannian manifolds. We propose a novel constructive approximation framework for…

Machine Learning · Statistics 2026-05-12 Zhongjie Shi , Wenjing Liao

Decision Transformer (DT) is an innovative algorithm leveraging recent advances of the transformer architecture in reinforcement learning (RL). However, a notable limitation of DT is its reliance on recalling trajectories from datasets,…

Machine Learning · Computer Science 2023-11-02 Yi Ma , Chenjun Xiao , Hebin Liang , Jianye Hao

Induction head mechanism is a part of the computational circuits for in-context learning (ICL) that enable large language models (LLMs) to adapt to new tasks without fine-tuning. Most existing work explains the training dynamics behind…

Computation and Language · Computer Science 2025-07-09 Shuo Wang , Issei Sato

Classical supervised learning produces unreliable models when training and target distributions differ, with most existing solutions requiring samples from the target domain. We propose a proactive approach which learns a relationship in…

Machine Learning · Statistics 2019-03-01 Adarsh Subbaswamy , Peter Schulam , Suchi Saria

Training data is always finite, making it unclear how to generalise to unseen situations. But, animals do generalise, wielding Occam's razor to select a parsimonious explanation of their observations. How they do this is called their…

Neurons and Cognition · Quantitative Biology 2023-07-20 William Dorrell , Maria Yuffa , Peter Latham

Modern machine learning requires system designers to specify aspects of the learning pipeline, such as losses, architectures, and optimizers. Meta-learning, or learning-to-learn, instead aims to learn those aspects, and promises to unlock…

Machine Learning · Computer Science 2024-01-10 Louis Kirsch , James Harrison , Jascha Sohl-Dickstein , Luke Metz

Can modifying the training data distribution guide optimizers toward solutions with improved generalization when training large language models (LLMs)? In this work, we theoretically analyze an in-context linear regression model with…

Machine Learning · Computer Science 2026-02-03 Tushaar Gangavarapu , Jiping Li , Christopher Vattheuer , Zhangyang Wang , Baharan Mirzasoleiman

Prior research shows that differences in the early period of neural network training significantly impact the performance of in-distribution (ID) data of tasks. Yet, the implications of early learning dynamics on out-of-distribution (OOD)…

Machine Learning · Computer Science 2025-01-07 Chen Cecilia Liu , Iryna Gurevych

In real world scenarios, out-of-distribution (OOD) datasets may have a large distributional shift from training datasets. This phenomena generally occurs when a trained classifier is deployed on varying dynamic environments, which causes a…

Image and Video Processing · Electrical Eng. & Systems 2022-09-08 Harshita Boonlia , Tanmoy Dam , Md Meftahul Ferdaus , Sreenatha G. Anavatti , Ankan Mullick

Understanding the training dynamics of transformers is important to explain the impressive capabilities behind large language models. In this work, we study the dynamics of training a shallow transformer on a task of recognizing…

Machine Learning · Computer Science 2024-10-15 Hongru Yang , Bhavya Kailkhura , Zhangyang Wang , Yingbin Liang

Next to scaling considerations, architectural design choices profoundly shape the solution space of transformers. In this work, we analyze the solutions simple transformer blocks implement when tackling the histogram task: counting items in…

Machine Learning · Computer Science 2025-11-13 Freya Behrens , Luca Biggio , Lenka Zdeborová
‹ Prev 1 3 4 5 6 7 10 Next ›