paper-with-me

Papers

How Does Overparameterization Affect Features?

2024-07-01 · Ahmet Cagri Duzgun, Samy Jelassi, Yuanzhi Li

Overparameterization, the condition where models have more parameters than necessary to fit their training loss, is a crucial factor for the success of deep learning. However, the characteristics of the features learned by overparameterized networks are not well understood. In this work, we explore this question by comparing models with the same architecture but different widths. We first examine the expressivity of the features of these models, and show that the feature space of overparameterized networks cannot be spanned by concatenating many underparameterized features, and vice versa. This reveals that both overparameterized and underparameterized networks acquire some distinctive features. We then evaluate the performance of these models, and find that overparameterized networks outperform underparameterized networks, even when many of the latter are concatenated. We corroborate these findings using a VGG-16 and ResNet18 on CIFAR-10 and a Transformer on the MNLI classification dataset. Finally, we propose a toy setting to explain how overparameterized networks can learn some important features that the underparamaterized networks cannot learn.

📄 PDF Abstract BibTeX arXiv:2407.00968

Code (0)

등록된 구현이 없습니다.

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Multi-Head Attention 설명 없음
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
VGG-16 설명 없음
BPE Byte Pair Encoding, or BPE, is a subword segmentation algorithm that encodes rare and unknown words as sequences of subword units. The intuition is that various word…
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…

Similar Papers 제목 키워드 기반

How does overparametrization affect performance on minority groups?

2022-06-07 · Subha Maity, Saptarshi Roy, Songkai Xue, Mikhail Yurochkin 외

The benefits of overparameterization for the overall performance of modern machine learning (ML) models are well known. However, the effect of overparameterization at a more granular level of data subgroups is less under…

regression

The Joint Effect of Task Similarity and Overparameterization on Catastrophic Forgetting -- An Analytical Model

2024-01-23 · Daniel Goldfarb, Itay Evron, Nir Weinberger, Daniel Soudry 외

In continual learning, catastrophic forgetting is affected by multiple aspects of the tasks. Previous works have analyzed separately how forgetting is affected by either task similarity or overparameterization. In contra…

Continual Learningregression

Critical Influence of Overparameterization on Sharpness-aware Minimization

2023-11-29 · Sungbin Shin, Dongyeop Lee, Maksym Andriushchenko, Namhoon Lee

Sharpness-Aware Minimization (SAM) has attracted considerable attention for its effectiveness in improving generalization in deep neural network training by explicitly minimizing sharpness in the loss landscape. Its succ…

Attribute

Memorization in Deep Neural Networks: Does the Loss Function matter?

2021-07-21 · Deep Patel, P. S. Sastry

Deep Neural Networks, often owing to the overparameterization, are shown to be capable of exactly memorizing even randomly labelled data. Empirical studies have also shown that none of the standard regularization techniq…

Memorization

Overparameterization hurts worst-group accuracy with spurious correlations

2020-01-01 · ICML 2020 1 · Shiori Sagawa, aditi raghunathan, Pang Wei Koh, Percy Liang

Increasing model capacity well beyond the point of zero training error has been observed to improve average test accuracy. However, such overparameterized models have been recently shown to obtain low worst-group accurac…