Federated reinforcement learning aims to promote training efficiency or improve policy quality through information interaction with privacy protection. Existing federated reinforcement learning methods rarely utilize the structure of reinforcement learning algorithms while are limiting to specific scenarios or algorithms. We propose a general federated reinforcement learning framework FRS, which employs reward shaping as the federated information shared among different clients with different tasks to promote each client’s training speed and policy quality. The federated reward shaping is implicitly learned by average state value information of all clients to protect each client’s task privacy as the real trajectory is anonymous. Experiments on the GridWorld environment show that FRS can algorithm-independently improve the policy quality and promote training speed with protecting each client’s privacy.
Background: Epidemiological studies have demonstrated that periodontitis is an independent risk factor for chronic obstructive pulmonary disease (COPD). However, the mechanism underlying the association between these two diseases remains unclear. The lung microbiota shares similarities with the oral microbiota, and there is growing evidence to suggest that the lung microbiome could play a role in the pathogenesis of COPD. This study aimed to investigate whether periodontal pathogens could contribute to the pathogenesis of COPD in a mouse model. Methods: We established mouse models with oral infection by typical periodontal pathogens, porphyromonas gingivalis (Pg group) or fusobacterium nucleatum (Fn group), over a three-month period. Mice that did not receive oral infection were set as the control group (C group). We assessed the level of alveolar bone resorption, lung function, and histological changes in the lungs of the mice. Additionally, we measured the levels of inflammatory factors and tissue damage associated factors in the lung tissues. Results: Lung function indices, including airway resistance, peak inspiratory/expiratory flow and expiratory flow-50%, were significantly reduced in the Fn group compared to the C group. Additionally, histological examination revealed an increased number of inflammatory cells and bullae formation in the lung tissue sections of the Fn group. Meanwhile, levels of inflammatory factors such as IL-1β, IL-6, IFN-γ, and TNF-α, as well as tissue damage associated factors like matrix metalloproteinase-8 and neutrophil elastase, were significantly elevated in the lung tissue of the Fn group in comparison to the C group. The Pg group also showed similar but milder lung changes compared to the Fn group. Pg or Fn could be detected in the lungs of both oral infected groups. Conclusion: The results indicated that oral periodontal pathogens infection could induce COPD-like lung changes in mice, and they may play a biological role in the association between periodontitis and COPD.
12 March 2024 in IEEE Transactions on Services Computing
Federated Learning (FL) is a promising paradigm for mining massive data while respecting users' privacy. However, the deployment of FL on resource-constrained edge devices remains elusive due to its high resource demand. In this paper, unlike existing works that use expensive dense models, we propose to utilize dynamic sparse training in FL and design a novel sparse-to-sparse FL framework, named as POP-FL. The framework can reduce both computation and communication overheads while maintaining the performance of the global model. Specifically, POP-FL partitions massive clients into groups and performs parallel parameter exploration, i.e., Parallel Over-Parameterization , over the collaboration between these groups. This exploration can greatly improve the expressibility and generalizability of sparse training in FL (especially for extreme sparsity levels) through reliably covering sufficient parameters and dynamically updating the global sparse network's structure during the training process. Experimental results show that compared with existing sparse-to-sparse training methods in both iid and non-iid data distribution, POP-FL achieves the best inference accuracy on various representative networks.
Graph neural networks (GNNs) are effective for structured data analysis but face reduced learning accuracy due to noisy connections and the necessity for explicit graph structures and labels. This requirement constrains their usability in diverse graph-based applications. In order to address these issues, considerable research has been directed toward graph structure learning that aims to denoise graph structures concurrently and refine GNN parameters. However, existing graph structure learning approaches encounter several challenges, including dependence on label information, underperformance of learning algorithms, insufficient data augmentation methods, and limitations in performing downstream tasks. We propose Uogtag, an unsupervised graph structure learning framework to address these challenges. Uogtag optimizes graph topology through the selection of suitable graph learners for the input data and incorporates contrastive learning with adaptive data augmentation, enhancing the learning and applicability of graph structures for downstream tasks. Comprehensive experiments on various real-world datasets demonstrate Uogtag’s efficacy in managing noisy graphs and label scarcity.