A Q- Learning based Routing Optimization Model in a Software Defined Network

Wang, Yupeng; Zhou, Xin; Fan, Xinyue

doi:10.1109/euc53437.2021.00029

Cited by 4 publications

(3 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…), it significantly changes the confidence of features, accelerates the convergence speed of the model, but the obtained weight values are not optimal. Experimental results have shown that when β is 1.45 (or 1.30), the Tight Binding mechanism [35] can better capture the spatial relationships and posture information between feature vectors, thereby improving the accuracy of tasks such as skin lesion diagnosis.…”

Section: B Backbone Networkmentioning

confidence: 99%

A Novel Skin Cancer Assisted Diagnosis Method based on Capsule Networks with CBAM

Lan¹,

Cai²,

Zhu³

et al. 2023

Preprint

View full text Add to dashboard Cite

HAM10000 consists of 10015 dermatoscopic images with a size of 450x600. There are seven types of skin lesions: actinic keratosis/intraepithelial carcinoma (AKIEC), basal cell carcinoma (BCC), benign keratosis (BKL), dermatofibroma (DF), melanoma (MEL), melanocytic nevi (NV), and vascular lesions (VASC). Dataset： https://www.kaggle.com/kmader/skin-cancer-mnist-ham10000 https://challenge.isic-archive.com/data/#2018 https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/DBW86T https://aistudio.baidu.com/aistudio/datasetdetail/218024 (ours)

show abstract

Section: B Backbone Networkmentioning

confidence: 99%

A Novel Skin Cancer Assisted Diagnosis Method based on Capsule Networks with CBAM

Lan¹,

Cai²,

Zhu³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…This includes different routing mechanisms, such as EM-Routing (Hinton, Sabour, and Frosst 2018), Self-Routing (Hahn, Pyeon, and Kim 2019), Variational Bayes Routing (Ribeiro, Leontidis, and Kollias 2020), Receptor Skeleton (Chen et al 2021) and attention-based routing (Ahmed and Torresani 2019;Tsai et al 2020;Mazzia, Salvetti, and Chiaberge 2021;Gu 2021). Wang and Liu (2018) reframed the routing algorithm in (Sabour, Frosst, and Hinton 2017) as an optimization problem, and Rawlinson, Ahmed, and Kowadlo (2018) introduced an unsupervised learning scheme for CapsNets. Other work replaced the capsule vector representations by matrices (Hinton, Sabour, and Frosst 2018) or tensors (Rajasegaran et al 2019), or added classic ConvNet features to the general routing mechanisms, such as dropout (Xiang et al 2018) or skipconnections (Rajasegaran et al 2019).…”

Section: Related Workmentioning

confidence: 99%

Why Capsule Neural Networks Do Not Scale: Challenging the Dynamic Parse-Tree Assumption

Mitterreiter

Koch

Giesen

et al. 2023

AAAI

View full text Add to dashboard Cite

Capsule neural networks replace simple, scalar-valued neurons with vector-valued capsules. They are motivated by the pattern recognition system in the human brain, where complex objects are decomposed into a hierarchy of simpler object parts. Such a hierarchy is referred to as a parse-tree. Conceptually, capsule neural networks have been defined to mimic this behavior. The capsule neural network (CapsNet), by Sabour, Frosst, and Hinton, is the first actual implementation of the conceptual idea of capsule neural networks. CapsNets achieved state-of-the-art performance on simple image recognition tasks with fewer parameters and greater robustness to affine transformations than comparable approaches. This sparked extensive follow-up research. However, despite major efforts, no work was able to scale the CapsNet architecture to more reasonable-sized datasets. Here, we provide a reason for this failure and argue that it is most likely not possible to scale CapsNets beyond toy examples. In particular, we show that the concept of a parse-tree, the main idea behind capsule neuronal networks, is not present in CapsNets. We also show theoretically and experimentally that CapsNets suffer from a vanishing gradient problem that results in the starvation of many capsules during training.

show abstract

“…Based on the performance research of the rotor-side converter of a doublyfed induction generator in a wind energy conversion system, Zhou et al ( 2018) [22] conducted corresponding comparative experiments on intelligent controllers based on Q-learning and dynamic fuzzy Q-learning, and the simulation results indicated that the intelligent controller based on Q-learning presented better performance. Regarding the problem of software-defined network routing optimization, Wang et al (2021) [23] not only proposed a routing algorithm based on Q-learning but also obtained the routing strategy and optimized the transmission delay, thus achieving the effect of ensuring user QoS.…”

Section: Model Solving Algorithm Based On Q-learningmentioning

confidence: 99%

Dynamic Resource Allocation Strategy of Multi-Objective Fuzzy Optimization Based on Markov Decision Process

Zhang,

Yang,

2023

IEEE Access

View full text Add to dashboard Cite

With the limited amount of national defense resources, the dynamic allocation of resources in overall project construction is conducive to improving the efficiency of project construction. In this regard, this paper constructs a dynamic resource allocation model of multi-objective and multi-stage fuzzy optimization based on the Markov decision process (MDP). On this basis, this paper further adopts the Q-learning algorithm to seek strategy optimization. At last, this paper determines the resource allocation strategy that can optimize the total benefit of construction projects through the example analysis. Relevant results reveal that the dynamic resource allocation model of multi-objective and multi-stage fuzzy optimization based on MDP can achieve the effect of optimizing the project construction objectives and maximizing the construction output.

show abstract

A Q- Learning based Routing Optimization Model in a Software Defined Network

Cited by 4 publications

References 15 publications

A Novel Skin Cancer Assisted Diagnosis Method based on Capsule Networks with CBAM

A Novel Skin Cancer Assisted Diagnosis Method based on Capsule Networks with CBAM

Why Capsule Neural Networks Do Not Scale: Challenging the Dynamic Parse-Tree Assumption

Dynamic Resource Allocation Strategy of Multi-Objective Fuzzy Optimization Based on Markov Decision Process

Contact Info

Product

Resources

About