One Bit Aggregation for Federated Edge Learning With Reconfigurable Intelligent Surface: Analysis and Optimization

Li, Heju; Wang, Rui; Zhang, Wei; Wu, Jun

doi:10.1109/twc.2022.3198881

Cited by 5 publications

(2 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[8] adopts multiple RISs to further enhance model uploading and solves a MSE minimization problem. In [9], RIS is applied to aid a one-bit communication FL system. [10] jointly allocates communication and computation resources to minimize average power consumption in an RIS-assisted OTA-FL system.…”

Section: Introductionmentioning

confidence: 99%

ROAR-Fed: RIS-Assisted Over-the-Air Adaptive Resource Allocation for Federated Learning

Mao,

Yener

2023

ICC 2023 - IEEE International Conference on Communications

View full text Add to dashboard Cite

Over-the-air federated learning (OTA-FL) integrates communication and model aggregation by exploiting the innate superposition property of wireless channels. The approach renders bandwidth efficient learning, but requires care in handling the wireless physical layer impairments. In this paper, federated edge learning is considered for a network that is heterogeneous with respect to client (edge node) data set distributions and individual client resources, under a general non-convex learning objective. We augment the wireless OTA-FL system with a Reconfigurable Intelligent Surface (RIS) to enable a propagation environment with improved learning performance in a realistic time varying physical layer. Our approach is a cross-layer perspective that jointly optimizes communication, computation and learning resources, in this general heterogeneous setting. We adapt the local computation steps and transmission power of the clients in conjunction with the RIS phase shifts. The resulting joint communication and learning algorithm, RIS-assisted Over-the-air Adaptive Resource Allocation for Federated learning (ROAR-Fed) is shown to be convergent in this general setting. Numerical results demonstrate the effectiveness of ROAR-Fed under heterogeneous (non i.i.d.) data and imperfect CSI, indicating the advantage of RIS assisted learning in this general set up.

show abstract

Section: Introductionmentioning

confidence: 99%

ROAR-Fed: RIS-Assisted Over-the-Air Adaptive Resource Allocation for Federated Learning

Mao,

Yener

2023

ICC 2023 - IEEE International Conference on Communications

View full text Add to dashboard Cite

show abstract

“…[1][2] By incorporating PIN diodes, [3][4][5][6][7][8] varactor diode, [9][10][11][12] graphene, [13] liquid crystal [14] and other components, [15][16][17][18] multi-parameters, [19][20][21][22][23][24][25] and multi-dimensional [3,26] modulation of electromagnetic waves are realized. Meta-devices such as reconfigurable intelligence surface (RIS), [27][28] phased array antennas, [29] and dynamic frequency selective surfaces [30] are constructed based on excellent electromagnetic modulation characteristics. However, most previous reconfigurable meta-devices work in 1D direction.…”

Section: Introductionmentioning

confidence: 99%

Active Matrix Integrated Metasurface for 2D Electromagnetic Modulation

Zhang,

et al. 2023

Advanced Optical Materials

View full text Add to dashboard Cite

Reconfigurable metasurfaces have attracted widespread attention owing to their dynamic, multifunctional, and intelligent manipulation ability of electromagnetic waves. However, most previous reconfigurable meta‐devices work in one‐dimension (1D) due to the complex electrical wires and control circuits for two‐dimensional (2D) modulation. Here, an active matrix circuit composed of a MOSFET chip and a storage capacitor is introduced into the meta‐atom design to realize 2D control of reconfigurable metasurfaces. The number of the control circuits is reduced from M × N to M + N by applying sequential voltage and selecting the row/column. As proof of the scheme, a 10 × 10 active matrix integrated reconfigurable metasurface operating in X‐band is experimentally demonstrated. 2D beam scanning and point focusing with focal lengths of 60 mm & 120 mm are verified through 2D phase modulation. The scheme for 2D modulation has great potential in the construction of high‐frequency large‐scale compact reconfigurable metasurfaces and promises unprecedented progress for wireless telecommunication and microwave detection.

show abstract