paper-with-me

홈 › Papers

Understanding and Improving the Role of Projection Head in Self-Supervised Learning

2022-12-22 · Kartik Gupta, Thalaiyasingam Ajanthan, Anton Van Den Hengel, Stephen Gould

Self-supervised learning (SSL) aims to produce useful feature representations without access to any human-labeled data annotations. Due to the success of recent SSL methods based on contrastive learning, such as SimCLR, this problem has gained popularity. Most current contrastive learning approaches append a parametrized projection head to the end of some backbone network to optimize the InfoNCE objective and then discard the learned projection head after training. This raises a fundamental question: Why is a learnable projection head required if we are to discard it after training? In this work, we first perform a systematic study on the behavior of SSL training focusing on the role of the projection head layers. By formulating the projection head as a parametric component for the InfoNCE objective rather than a part of the network, we present an alternative optimization scheme for training contrastive learning based SSL frameworks. Our experimental study on multiple image classification datasets demonstrates the effectiveness of the proposed approach over alternatives in the SSL literature.

📄 PDF Abstract BibTeX arXiv:2212.11491

Code (0)

등록된 구현이 없습니다.

Tasks

Contrastive Learningimage-classificationImage ClassificationSelf-Supervised Learning

Methods 이 논문이 사용한 방법론

Bitcoin Customer Service Number +1-833-534-1729 설명 없음
Average Pooling 설명 없음
Batch Normalization 설명 없음
Residual Block Residual Blocks are skip-connection blocks that learn residual functions with reference to the layer inputs, instead of learning unreferenced functions. They were introduced…
Global Average Pooling Global Average Pooling is a pooling operation designed to replace fully connected layers in classical CNNs. The idea is to generate one feature map for each corresponding…
1x1 Convolution A 1 x 1 Convolution is a convolution with some special properties in that it can be used for dimensionality reduction,…
Kaiming Initialization 설명 없음
Convolution A convolution is a type of matrix operation, consisting of a kernel, a small matrix of weights, that slides over input data performing element-wise multiplication with the…

Similar Papers 제목 키워드 기반

Deciphering the Projection Head: Representation Evaluation Self-supervised Learning

2023-01-28 · Jiajun Ma, Tianyang Hu, Wenjia Wang

Self-supervised learning (SSL) aims to learn intrinsic features without labels. Despite the diverse architectures of SSL methods, the projection head always plays an important role in improving the performance of the dow…

Self-Supervised Learning

Towards the Sparseness of Projection Head in Self-Supervised Learning

2023-07-18 · Zeen Song, Xingzhe Su, Jingyao Wang, Wenwen Qiang 외

In recent years, self-supervised learning (SSL) has emerged as a promising approach for extracting valuable representations from unlabeled data. One successful SSL method is contrastive learning, which aims to bring posi…

Contrastive LearningSelf-Supervised Learning

Retro: Reusing teacher projection head for efficient embedding distillation on Lightweight Models via Self-supervised Learning

2024-05-24 · Khanh-Binh Nguyen, Chae Jung Park

Self-supervised learning (SSL) is gaining attention for its ability to learn effective representations with large amounts of unlabeled data. Lightweight models can be distilled from larger self-supervised pre-trained mod…

Self-Supervised Learning

Projection Head is Secretly an Information Bottleneck

2025-03-01 · Zhuo Ouyang, Kaiwen Hu, Qi Zhang, Yifei Wang 외

Recently, contrastive learning has risen to be a promising paradigm for extracting meaningful data representations. Among various special designs, adding a projection head on top of the encoder during training and removi…

Contrastive Learning

A Probabilistic Model Behind Self-Supervised Learning

2024-02-02 · Alice Bizeul, Bernhard Schölkopf, Carl Allen

In self-supervised learning (SSL), representations are learned via an auxiliary task without annotated labels. A common task is to classify augmentations or different modalities of the data, which share semantic content …

modelRepresentation LearningSelf-Supervised Learning