Conference reference AIIoT in VIT PPT.pptx

AIIoT

1. Agenda Abstract Introduction Literature Survey Proposed Methodology Conclusion References AIIoT 2024 1

2. Abstract Federated learning (FL) has grown in popularity in all fields and analyzing large amounts of data from IoT devices. The privacy of IoT device input data is problematic due to its unreliability. The proposed method offers a secure aggregation mechanism using effectual additive secret allotment in the fog computing (FC) environment. The protocol should minimize communication latency and processing time. We deploy a fog node (FN) as intermediary units to deliver local request that help the cloud aggregate the sum of tasks. We designed a Fog-FL mechanism to ensure our proposed method is robust against dropped user tasks. Achieved an increment up to 8.5% of global accuracy as compared with reported methods. AIIoT 2024 2

3. Introduction Federated Learning (FL) allows IoT devices to develop AI models cooperatively without sharing local data. S tandard FL approaches have the following issues : Communication complexity (or) overhead caused by the constant transfer of information about models between cloud and the edge device ; also, poor connectivity results in latency issues. Sophisticated methods of calculation for training require high-performance devices. However, the IoT system's edge devices typically have limited energy, storage, and processing power resources. They depend mainly on a single centralized structure for each period. Such a centralized structure is susceptible to problems with single points of failure and bottlenecks, leading to an unsustainable global model AIIoT 2024 3

4. Introduction T he cloud may struggle to handle vast training data, particularly for privacy preserving deep learning techniques. This study makes the following research contributions: Using the Smart Decision Making module, we designed an early warning system for untrusted data and Fog Node We integrate the Federated Learning approach into the fog layer to ensure end-user data privacy. Determine an acceptable placement aggregation about the FL module in the Edge-Fog-Cloud layers. AIIoT 2024 4

5. Literature Survey Yang et al:- Suggested a task selection policy to calculate the time required for transmitting and training the model using all measured resources. Li et al:- Cloud-fog architecture was utilized to provide secure, privacy-preserving, distributed deep learning training . Saha et al:- S uggested a fog-based FL architecture for delay-sensitive data . FL occurs between the edge and fog layers, followed by heuristic global model aggregation at the central node. AIIoT 2024 5

6. Literature Survey Huang X et al:- Explored a method to optimize content caching and user experience in the IoT context. They provided the critical aspects of FL, Quality of Service ( QoS ), and Caching to reduce retrieval delays . Hasan BT et al:- Proposed a framework for FL that leverages the collaborative capabilities of edge/fog devices. Xu et al:- Introduced a multi-level split FL (SFL) framework that Combines cloud based processing with edge and fog computing layers. Then utilizes the Message Queuing Telemetry Transport (MQTT) protocol for geographically clustering devices and assigning them to edge/ FNs . AIIoT 2024 6

7. Proposed Methodology System model AIIoT 2024 7

8. Proposed Methodology The set of K-edge nodes is taken as IoT /user devices F -FNs are taken as base servers which is responsible for global and local aggregation within the fog network A cloud is the coordination node for global processing and monitoring of the system . The various phases are as follows: Selection of IoT/User devices. Configuration and local updates on inputs. Local aggregation in edge node. Global aggregation in fog nodes. Final reporting in cloud. AIIoT 2024 8

9. Proposed Methodology Fog-FL Time-based training model for honest fog node selection AIIoT 2024 9

10. Conclusion The Fog-FL framework was created and implemented for edge devices with geographically positioned FN s to gather for local data updates and globally aggregate nodes in this model layer. W e formulate a strategy for choosing the best FN to serve as a global aggregate node at each round . T he proposed FoG -FL outperforms other existing methods, which have an average loss function of less than 22% and a global accuracy of 58% in the initial round itself . Future research could focus on developing a secure aggregation algorithm for malicious clients with high input volumes. AIIoT 2024 10

11. References B. Premalatha , P. Prakasam , "Optimal Energy-efficient Resource Allocation and Fault Tolerance scheme for task offloading in IoT-FoG Computing Networks." Computer Networks, vol. 238, pp.110080, Jan 2024, doi : 10.1016/j.comnet.2023.110080 . H.Kim , J.Park , M.Bennis , and S.- L.Kim , “ BlockchainedOn -Device FederatedLearning ,” IEEE Communications Letters, vol.24, no.6, pp. 1279–1283,2019. T. Zhang and S. Mao, “An introduction to the federated learning standard,” GetMobile : Mobile Computing and Communications, vol. 25, no. 3, pp. 18–22, 2022. Y. J. Cho, J. Wang, and G. Joshi, “Client selec tion in federated learning: Convergence analysis and power-of-choice selection strategies,” arXiv preprint arXiv:2010.01243, 2020. H. Yang, J. Zhao, Z. Xiong , K.-Y. Lam, S. Sun, and L. Xiao, “Privacy-preserving federated learning for uav enabled networks: Learning-based joint scheduling and resource management,” IEEE Journal on Selected Areas in Communications, vol. 39, no. 10, pp. 3144–3159, 2021. Y. Li, H. Li, G. Xu, T. Xiang, X. Huang, R. Lu, “Toward Secure and Privacy-Preserving Distributed Deep Learning in Fog-Cloud Computing”, IEEE Internet of Things Journal, vol. 7, no. 12, pp. 11460–11472. 2020, doi:10. 1109/JIOT.2020.3012480. R. Saha , S. Misra , P. K. Deb, “ FogFL : Fog-Assisted Federated Learning for Resource-Constrained IoT Devices”, IEEE Internet of Things Journal, vol. 8, no. 10, pp. 8456–8463, 2021. doi:10.1109/JIOT.2020.3046509. S.S. Tripathy , S. Bebortta , C.L. Chowdhary , T. Mukherjee, S. Kim, J. Shafi , M.F. Ijaz , “ FedHealthFog : A federated learning-enabled approach towards healthcare analytics over fog computing platform”, Heliyon . 2024 Feb 16. X. Huang, Z. Chen, Q. Chen, J. Zhang, “Federated learning based qos -aware caching decisions in fog-enabled internet of things networks”, Digital Communications and Networks, vol. 9, no. 2, pp. 580-589, 2023. B.T. Hasan, A.K. Idrees , “Federated Learning for Iot /Edge/Fog Computing Systems”, arXiv preprint arXiv:2402.13029, Feb 2024. H. Xu, K.P. Seng, J. Smith, L.M. Ang , “Multi-Level Split Federated Learning for Large-Scale AIoT System Based on Smart Cities”, Future Internet, vol. 16, no. 3, 2024, https://doi.org/10.3390/fi16030082 . AIIoT 2024 11