Deep Meta Functionals for Shape Representation

Littwin, Gidi; Lior, Wolf

doi:10.48550/arxiv.1908.06277

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The notion of a hypernetwork has been used in several contexts such as object recognition [20] and generation of 3-D point clouds [21]. For example, hypernetworks have been used for 3D shape reconstruction [22] and to learn shared representations of images [23]. In the specific context of MIMO detection, the use of hypernetworks has been already proposed in [16] applied to the MM-Net.…”

Section: Hypernetwork In Machine Learning and Mimo Detectionmentioning

confidence: 99%

Robust MIMO Detection using Hypernetworks with Learned Regularizers

Zilberstein¹,

Dick²,

Doost-Mohammady³

et al. 2021

Preprint

View full text Add to dashboard Cite

Optimal symbol detection in multiple-input multiple-output (MIMO) systems is known to be an NP-hard problem. Recently, there has been a growing interest to get reasonably close to the optimal solution using neural networks while keeping the computational complexity in check. However, existing work based on deep learning shows that it is difficult to design a generic network that works well for a variety of channels. In this work, we propose a method that tries to strike a balance between symbol error rate (SER) performance and generality of channels. Our method is based on hypernetworks that generate the parameters of a neural network-based detector that works well on a specific channel. We propose a general framework by regularizing the training of the hypernetwork with some pre-trained instances of the channel-specific method. Through numerical experiments, we show that our proposed method yields high performance for a set of prespecified channel realizations while generalizing well to all channels drawn from a specific distribution.

show abstract