Incremental Weighted Ensemble for Data Streams With Concept Drift

Jiao, Botao; Guo, Yinan; Yang, Cuie; Pu, Jiayang; Zheng, Zhiji; Gong, Dunwei

doi:10.1109/tai.2022.3224416

Cited by 9 publications

(3 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Class distributions are generally imbalanced in real-world data, as IEEE TRANSACTIONS ON ARTIFICIAL INTELLIGENCE certain objects or patterns appear more frequently. Various works have been done to tackle the imbalanced distribution in recent years, which raises the importance of evaluating deep learning algorithms on imbalanced datasets [29][30][31][32]. To study the effectiveness of the proposed method for classifying imbalanced image datasets, we adopted one synthetic imbalanced dataset based on CIFAR10 and one real-world dataset, the SVHN dataset.…”

Section: ) Classification Performance On Imbalanced Datasetsmentioning

confidence: 99%

Supervised Local Training With Backward Links for Deep Neural Networks

Guo

Fouda

Eltawil

et al. 2024

IEEE Trans. Artif. Intell.

View full text Add to dashboard Cite

The restricted training pattern in the standard BP requires end-to-end error propagation, causing large memory costs and prohibiting model parallelization. Existing local training methods aim to resolve the training obstacles by completely cutting off the backward path between modules and isolating their gradients. These methods prevent information exchange between modules and result in inferior performance. This work proposes a novel local training algorithm, BackLink, which introduces intermodule backward dependency and facilitates information to flow backward along with the network. To preserve the computational advantage of local training, BackLink restricts the error propagation length within the module. Extensive experiments performed in various deep convolutional neural networks demonstrate that our method consistently improves the classification performance of local training algorithms over other methods. For example, our method can surpass the conventional greedy local training method by 6.45% in accuracy in ResNet32 classifying CIFAR100 and recent work by 2.58% in ResNet110 classifying STL-10 with much lower complexity, respectively. Analysis of computational costs reveals that small overheads are incurred in GPU memory costs and runtime on multiple GPUs. Our method can lead up to a 79% reduction in memory cost and 52% in simulation runtime in ResNet110 compared to the standard BP. Therefore, our method could create new opportunities for improving training algorithms towards better efficiency for real-time learning applications. Impact Statement-Deep learning models are generally bulky with abundant parameters, demanding long training time and excessive computing power. It poses a great challenge for deploying these models in real-world applications. This study looks into the intrinsic limitation of the standard training algorithm. The proposed method increases information flow in the algorithm and improves image classification accuracy over conventional methods. It can substantially reduce memory cost and hardware runtime by 50%. With the proposed method, an intelligent system can be readily deployed on mobile devices and function to meet practical demands. It also enables computing systems to reduce excessive energy waste and cut down carbon footprints for deep learning applications, such as traffic control and autonomous driving.

show abstract

Section: ) Classification Performance On Imbalanced Datasetsmentioning

confidence: 99%

Supervised Local Training With Backward Links for Deep Neural Networks

Guo

Fouda

Eltawil

et al. 2024

IEEE Trans. Artif. Intell.

View full text Add to dashboard Cite

show abstract

“…C. Zhao et al 's (23) fine-tuning ensemble models for ensemble learning can enhance the accuracy and adaptability of crop recommendation systems. Jiao B et al 's (24) incremental weighted ensemble for data streams is crucial for crop recommendation models to adapt to changing farming practices, climate conditions, and other factors. Akkem Y et al 's (17,18) smart farming using artificial intelligence and smart farming monitoring using ML and MLOps provide insights that can inform the development and enhancement of artificial intelligence techniques, including ensemble learning, for smart farming applications.…”

Section: Related Workmentioning

confidence: 99%

“…These systems have been pivotal in processing complex agronomic data to facilitate informed decision-making. While various machine learning-based methods have been employed, there is a crucial need to recognize the constraints of existing ML techniques used in crop recommendation systems when addressing the evolving complexities of this field (20)(21)(22)(23)(24) . Furthermore, despite the abundance of literature on ML applications in agriculture, there has been a lack of comprehensive consideration for advanced ensemble learning approaches and the power of Streamlit for handling current challenges in agricultural data analysis (5)(6)(7)11,12) .…”

Section: Introductionmentioning

confidence: 99%

Streamlit Application for Advanced Ensemble Learning Methods in Crop Recommendation Systems – A Review and Implementation

Akkem,

Kumar,

Varanasi

2023

IJST

View full text Add to dashboard Cite

Objectives: This article explores the integration of advanced ensemble machine learning methods within precision agriculture, aiming to enhance the reliability and practical utility of crop recommendation systems. The incorporation of the Streamlit framework in the development process underpins our objective to deliver a user-friendly tool that provides farmers and agricultural analysts with actionable insights. Methods: A thorough literature review of artificial intelligence applications in agriculture serves as the foundation of our study, with a strong emphasis placed on sophisticated ensemble learning techniques such as stacking, an ensemble of ensembles, and federated learning. The evaluation methodology entails a comparative analysis where these cutting-edge techniques are juxtaposed against standard machine learning benchmarks to ascertain their performance improvement. In addition to the conceptual analysis, we implement a crop recommendation system using the Streamlit framework, emphasizing usability and accessibility for end-users to interact with machine learning predictions based on their soil data. Findings: The empirical results demonstrate that our chosen advanced ensemble learning methods significantly improve predictive performance, recording up to a 15% accuracy increment over traditional machine learning algorithms. Their adaptability to varied agricultural datasets, coupled with robust privacypreserving properties, stand out. When deploying these methods in a practical Streamlit-based application, we note a marked increase of 20% in user efficiency, solidifying the system's crucial role in bolstering resilient crop management tactics. Novelty: This research pioneers the study of innovative ensemble learning techniques, married with Streamlit app development for an enhanced user experience in data-driven precision agriculture. Our findings emphasize the critical need for incorporating these advanced methodologies into realworld practices, fostering a significant paradigm shift in agricultural data analytics and management. The synergy between these powerful machine learning techniques and the Streamlit-built interactive interface represents a step https://www.indjst.org/ forward in translating complex computational analysis into practical, on-theground tools for agriculture professionals.

show abstract