Incremental Learning in Online Scenario

He, Jiangpeng; Mao, Runyu; Shao, Zeman; Zhu, Fengqing

doi:10.1109/cvpr42600.2020.01394

Cited by 121 publications

(88 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Most data replay-based incremental learning methods [22], [43], [44] follow the iCaRL experiment benchmark protocol to arrange classes and select exemplars. In ScaIL [45], the experimental results show that exemplar selection based on herding can improve performance.…”

Section: B Main Componentsmentioning

confidence: 99%

Incremental SAR Automatic Target Recognition With Error Correction and High Plasticity

Tang

Xiang

Zhang

et al. 2022

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Synthetic aperture radar automatic target recognition (SAR ATR) uses computer processing capabilities to infer the classes of the targets without human intervention. For SAR ATR, deep learning gradually emerges as a powerful tool and achieves promising performance. However, it faces serious challenges of how to deal with incremental recognition scenarios. The existing deep learning-based SAR ATR methods usually predefine the total number of recognition classes. In realistic applications, the new tasks/classes will be added continuously. If all old data are stored and mixed with newly added data to update the model, the storage pressure and time consumption make the application infeasible. In this article, the high plastic error correction incremental learning (HPecIL) is proposed to address the model degradation and plasticity decline in the incremental scenario. Multiple optimal models trained on old tasks are used to correct accumulative errors and alleviate model degradation. Moreover, the sharp data distribution shift due to newly added data can also result in the model underperforming. A classbalanced training batch is constructed to deal with the issue of unbalanced data distribution. To make a trade-off between model stability and model plasticity, low-effect nodes in the model are removed to boost the efficiency of model update. The proposed HPecIL outperforms the other state-of-the-art methods in incremental recognition scenarios. The experimental results demonstrate the effectiveness of the proposed method.

show abstract

Section: B Main Componentsmentioning

confidence: 99%

Incremental SAR Automatic Target Recognition With Error Correction and High Plasticity

Tang

Xiang

Zhang

et al. 2022

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

show abstract

“…On food recognition, we developed a two-step approach for food localization and hierarchical food classification using Convolutional Neural Networks (CNNs) to reduce prediction error for visually similar foods [23], and continual learning in the challenging online learning scenario that is further bounded by run-time and limited data [18,20]. On image segmentation, we developed class-agnostic method using a pair of eating scene images to find the salient missing objects without prior information about the food class [41] and developed weakly supervised and efficient superpixel based methods [38,39].…”

Section: Food Image Analysismentioning

confidence: 99%

An Integrated System for Mobile Image-Based Dietary Assessment

Shao¹,

Han²,

He³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Accurate assessment of dietary intake requires improved tools to overcome limitations of current methods including user burden and measurement error. Emerging technologies such as imagebased approaches using advanced machine learning techniques coupled with widely available mobile devices present new opportunities to improve the accuracy of dietary assessment that is costeffective, convenient and timely. However, the quality and quantity of datasets are essential for achieving good performance for automated image analysis. Building a large image dataset with high quality groundtruth annotation is a challenging problem, especially for food images as the associated nutrition information needs to be provided or verified by trained dietitians with domain knowledge. In this paper, we present the design and development of a mobile, image-based dietary assessment system to capture and analyze dietary intake, which has been deployed in both controlled-feeding and community-dwelling dietary studies. Our system is capable of collecting high quality food images in naturalistic settings and provides groundtruth annotations for developing new computational approaches.

show abstract

“…Still, due to the unpredictable nature of COVID-19, the modeling process should be updated as often as possible with the newly collected data -as insights contained in the novel data may introduce information that may improve the models, enabling a higher precision. This may be achieved either using incremental learning paradigm (also known as online learning) [18] in which the model is retrained using newly acquired data or through the full retraining of the models with the entire dataset (including old, and newly acquired data) [19].…”

Section: Introductionmentioning

confidence: 99%

Automated Pipeline for Continual Data Gathering and Retraining of the Machine Learning-Based COVID-19 Spread Models

Šegota¹,

Lorencin²,

Anđelić³

et al. 2021

EAI Endorsed Transactions on Bioengineering and Bioinformatics

View full text Add to dashboard Cite

INTRODUCTION:The development of epidemiological curve models is one of the key factors in the combat of epidemiological diseases such as COVID-19. OBJECTIVES: The goal of this paper is to develop a system for automatic training and testing of AI-based regressive models of epidemiological curves using public data, which involves automating the data acquisition and speeding up the training of the models. METHODS: The research applies Multilayer Perceptron (MLP) for the creation of models, implemented within a system for automatic data fetching and training, and evaluated using the coefficient of determination (R 2 ). Training time is lowered through the application of data filtering and simplifying the model selection. RESULTS: The developed system can train high precision models rapidly, allowing for quick model delivery All trained models achieve scores which are higher than 0.95. CONCLUSION: The results show that the development of a quick COVID-19 spread modeling system is possible.

show abstract

Incremental Learning in Online Scenario

Cited by 121 publications

References 7 publications

Incremental SAR Automatic Target Recognition With Error Correction and High Plasticity

Incremental SAR Automatic Target Recognition With Error Correction and High Plasticity

An Integrated System for Mobile Image-Based Dietary Assessment

Automated Pipeline for Continual Data Gathering and Retraining of the Machine Learning-Based COVID-19 Spread Models

Contact Info

Product

Resources

About