On Improving the Algorithm-, Model-, and Data- Efficiency of Self-Supervised Learning
- URL: http://arxiv.org/abs/2404.19289v1
- Date: Tue, 30 Apr 2024 06:39:04 GMT
- Title: On Improving the Algorithm-, Model-, and Data- Efficiency of Self-Supervised Learning
- Authors: Yun-Hao Cao, Jianxin Wu,
- Abstract summary: We propose an efficient single-branch SSL method based on non-parametric instance discrimination.
We also propose a novel self-distillation loss that minimizes the KL divergence between the probability distribution and its square root version.
- Score: 18.318758111829386
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: Self-supervised learning (SSL) has developed rapidly in recent years. However, most of the mainstream methods are computationally expensive and rely on two (or more) augmentations for each image to construct positive pairs. Moreover, they mainly focus on large models and large-scale datasets, which lack flexibility and feasibility in many practical applications. In this paper, we propose an efficient single-branch SSL method based on non-parametric instance discrimination, aiming to improve the algorithm, model, and data efficiency of SSL. By analyzing the gradient formula, we correct the update rule of the memory bank with improved performance. We further propose a novel self-distillation loss that minimizes the KL divergence between the probability distribution and its square root version. We show that this alleviates the infrequent updating problem in instance discrimination and greatly accelerates convergence. We systematically compare the training overhead and performance of different methods in different scales of data, and under different backbones. Experimental results show that our method outperforms various baselines with significantly less overhead, and is especially effective for limited amounts of data and small models.
Related papers
- Linearly Convergent Mixup Learning [0.0]
We present two novel algorithms that extend to a broader range of binary classification models.
Unlike gradient-based approaches, our algorithms do not require hyper parameters like learning rates, simplifying their implementation and optimization.
Our algorithms achieve faster convergence to the optimal solution compared to descent gradient approaches, and that mixup data augmentation consistently improves the predictive performance across various loss functions.
arXiv Detail & Related papers (2025-01-14T02:33:40Z) - SeMi: When Imbalanced Semi-Supervised Learning Meets Mining Hard Examples [54.760757107700755]
Semi-Supervised Learning (SSL) can leverage abundant unlabeled data to boost model performance.
The class-imbalanced data distribution in real-world scenarios poses great challenges to SSL, resulting in performance degradation.
We propose a method that enhances the performance of Imbalanced Semi-Supervised Learning by Mining Hard Examples (SeMi)
arXiv Detail & Related papers (2025-01-10T14:35:16Z) - A Bayesian Approach to Data Point Selection [24.98069363998565]
Data point selection (DPS) is becoming a critical topic in deep learning.
Existing approaches to DPS are predominantly based on a bi-level optimisation (BLO) formulation.
We propose a novel Bayesian approach to DPS.
arXiv Detail & Related papers (2024-11-06T09:04:13Z) - Augmentations vs Algorithms: What Works in Self-Supervised Learning [9.194402355758164]
We study the relative effects of data augmentations, pretraining algorithms, and model architectures in Self-Supervised Learning (SSL)
We propose a new framework which unifies many seemingly disparate SSL methods into a single shared template.
arXiv Detail & Related papers (2024-03-08T23:42:06Z) - Improved Distribution Matching for Dataset Condensation [91.55972945798531]
We propose a novel dataset condensation method based on distribution matching.
Our simple yet effective method outperforms most previous optimization-oriented methods with much fewer computational resources.
arXiv Detail & Related papers (2023-07-19T04:07:33Z) - Revisiting Consistency Regularization for Semi-Supervised Learning [80.28461584135967]
We propose an improved consistency regularization framework by a simple yet effective technique, FeatDistLoss.
Experimental results show that our model defines a new state of the art for various datasets and settings.
arXiv Detail & Related papers (2021-12-10T20:46:13Z) - Simple Stochastic and Online Gradient DescentAlgorithms for Pairwise
Learning [65.54757265434465]
Pairwise learning refers to learning tasks where the loss function depends on a pair instances.
Online descent (OGD) is a popular approach to handle streaming data in pairwise learning.
In this paper, we propose simple and online descent to methods for pairwise learning.
arXiv Detail & Related papers (2021-11-23T18:10:48Z) - Attentional-Biased Stochastic Gradient Descent [74.49926199036481]
We present a provable method (named ABSGD) for addressing the data imbalance or label noise problem in deep learning.
Our method is a simple modification to momentum SGD where we assign an individual importance weight to each sample in the mini-batch.
ABSGD is flexible enough to combine with other robust losses without any additional cost.
arXiv Detail & Related papers (2020-12-13T03:41:52Z) - End-to-End Training of CNN Ensembles for Person Re-Identification [0.0]
We propose an end-to-end ensemble method for person re-identification (ReID) to address the problem of overfitting in discriminative models.
Our proposed ensemble learning framework produces several diverse and accurate base learners in a single DenseNet.
Experiments on several benchmark datasets demonstrate that our method achieves state-of-the-art results.
arXiv Detail & Related papers (2020-10-03T12:40:13Z) - Extrapolation for Large-batch Training in Deep Learning [72.61259487233214]
We show that a host of variations can be covered in a unified framework that we propose.
We prove the convergence of this novel scheme and rigorously evaluate its empirical performance on ResNet, LSTM, and Transformer.
arXiv Detail & Related papers (2020-06-10T08:22:41Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.