English
Related papers

Related papers: A Closer Look at Model Collapse: From a Generaliza…

200 papers

Continual learning (CL), which aims to learn a sequence of tasks, has attracted significant recent attention. However, most work has focused on the experimental performance of CL, and theoretical studies of CL are still limited. In…

Machine Learning · Computer Science 2023-02-14 Sen Lin , Peizhong Ju , Yingbin Liang , Ness Shroff

As access to high-quality, domain-specific data grows increasingly scarce, multi-epoch training has become a practical strategy for adapting large language models (LLMs). However, autoregressive models often suffer from performance…

Computation and Language · Computer Science 2025-12-30 Jiapeng Wang , Yiwen Hu , Yanzipeng Gao , Haoyu Wang , Shuo Wang , Hongyu Lu , Jiaxin Mao , Wayne Xin Zhao , Junyi Li , Xiao Zhang

Recent work on diffusion models proposed that they operate in two regimes: memorization, in which models reproduce their training data, and generalization, in which they generate novel samples. While this has been tested in high-noise…

Computer Vision and Pattern Recognition · Computer Science 2025-06-10 Elizabeth Pavlova , Xue-Xin Wei

When language models (LMs) are trained to forget (or "unlearn'') a skill, how precisely does their behavior change? We study the behavior of transformer LMs in which tasks have been forgotten via fine-tuning on randomized labels. Such LMs…

Machine Learning · Computer Science 2024-09-05 Eric Zhang , Leshem Chosen , Jacob Andreas

The training dynamics of deep neural networks often defy expectations, even as these models form the foundation of modern machine learning. Two prominent examples are grokking, where test performance improves abruptly long after the…

Machine Learning · Computer Science 2026-01-28 Keitaro Sakamoto , Issei Sato

When do language diffusion models memorize their training data, and how to quantitatively assess their true generative regime? We address these questions by showing that Uniform-based Discrete Diffusion Models (UDDMs) fundamentally behave…

Machine Learning · Computer Science 2026-04-30 Bao Pham , Mohammed J. Zaki , Luca Ambrogioni , Dmitry Krotov , Matteo Negri

Overfitting negatively impacts the generalization ability of deep neural networks (DNNs) in both natural and adversarial training. Existing methods struggle to consistently address different types of overfitting, typically designing…

Machine Learning · Computer Science 2024-09-17 Runqi Lin , Chaojian Yu , Bo Han , Tongliang Liu

Synthetically-generated data plays an increasingly larger role in training large language models. However, while synthetic data has been found to be useful, studies have also shown that without proper curation it can cause LLM performance…

Machine Learning · Computer Science 2025-12-02 Kareem Amin , Sara Babakniya , Alex Bie , Weiwei Kong , Umar Syed , Sergei Vassilvitskii

We first study the generalization error of models that use a fixed feature representation (frozen intermediate layers) followed by a trainable readout layer. This setting encompasses a range of architectures, from deep random-feature models…

Statistics Theory · Mathematics 2025-11-10 Yessin Moakher , Malik Tiomoko , Cosme Louart , Zhenyu Liao

The past few years have witnessed substantial advances in image generation powered by diffusion models. However, it was shown that diffusion models are susceptible to training data memorization, raising significant concerns regarding…

Cryptography and Security · Computer Science 2025-08-01 Zhe Ma , Qingming Li , Xuhong Zhang , Tianyu Du , Ruixiao Lin , Zonghui Wang , Shouling Ji , Wenzhi Chen

It is known that recursive training from generative models can lead to the so called `collapse' of the simulated probability distribution. This note shows that one in fact gets two different asymptotic behaviours depending on whether an…

Probability · Mathematics 2025-09-30 Vivek Shripad Borkar

The proliferation of AI-generated content online has fueled concerns over \emph{model collapse}, a degradation in future generative models' performance when trained on synthetic data generated by earlier models. Industry leaders, premier…

Machine Learning · Computer Science 2025-03-19 Rylan Schaeffer , Joshua Kazdan , Alvan Caleb Arulandu , Sanmi Koyejo

Loss of plasticity is one of the main challenges in continual learning with deep neural networks, where neural networks trained via backpropagation gradually lose their ability to adapt to new tasks and perform significantly worse than…

Machine Learning · Computer Science 2025-03-27 Jiuqi Wang , Rohan Chandra , Shangtong Zhang

Class-incremental learning aims to learn new classes in an incremental fashion without forgetting the previously learned ones. Several research works have shown how additional data can be used by incremental models to help mitigate…

Machine Learning · Computer Science 2023-10-11 Quentin Jodelet , Xin Liu , Yin Jun Phua , Tsuyoshi Murata

We aim to understand grokking, a phenomenon where models generalize long after overfitting their training set. We present both a microscopic analysis anchored by an effective theory and a macroscopic analysis of phase diagrams describing…

Machine Learning · Computer Science 2022-10-17 Ziming Liu , Ouail Kitouni , Niklas Nolte , Eric J. Michaud , Max Tegmark , Mike Williams

Will further scaling up of machine learning models continue to bring success? A significant challenge in answering this question lies in understanding generalization gap, which is the impact of overfitting. Understanding generalization gap…

Machine Learning · Statistics 2026-05-18 Atsushi Suzuki , Jing Wang

In-context learning enables transformer models to generalize to new tasks based solely on input prompts, without any need for weight updates. However, existing training paradigms typically rely on large, unstructured datasets that are…

Catastrophic forgetting (CF) happens whenever a neural network overwrites past knowledge while being trained on new tasks. Common techniques to handle CF include regularization of the weights (using, e.g., their importance on past tasks),…

Machine Learning · Computer Science 2022-12-29 Jary Pomponi , Simone Scardapane , Aurelio Uncini

Training diffusion models requires large datasets. However, acquiring large volumes of high-quality data can be challenging, for example, collecting large numbers of high-resolution images and long videos. On the other hand, there are many…

Computer Vision and Pattern Recognition · Computer Science 2025-05-20 Xudong Ma

Recent research has highlighted the importance of dataset size in scaling language models. However, large language models (LLMs) are notoriously token-hungry during pre-training, and high-quality text data on the web is approaching its…

Machine Learning · Computer Science 2023-10-10 Fuzhao Xue , Yao Fu , Wangchunshu Zhou , Zangwei Zheng , Yang You