A Meta Learner Autoencoder for Channel State Information Feedback in Massive MIMO Systems

Tolba, Bassant; El‐Malek, Ahmed H. Abd; Abo‐Zahhad, Mohammed; Elsabrouty, Maha

doi:10.1109/ict52184.2021.9511541

Cited by 5 publications

(2 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, plenty of training data of target scenario is quite impractical for deployment due to the expense and long-time training and collecting data. Meta-learning is utilized for CSI feedback in [8] and [9], where the model is initialized by the meta model obtained in meta training phase with massive CSI samples corresponding to multiple various scenarios, and then achieves quick convergence with small amount of CSI data in a new target scenario. However, the above meta-learning based solutions still require massive collected data for the meta training phase.…”

Section: Introductionmentioning

confidence: 99%

A Knowledge-Driven Meta-Learning Method for CSI Feedback

Han¹,

Tian²,

Li³

et al. 2023

Preprint

View full text Add to dashboard Cite

Accurate and effective channel state information (CSI) feedback is a key technology for massive multiple-input and multiple-output (MIMO) systems. Recently, deep learning (DL) has been introduced to enhance CSI feedback in massive MIMO application, where the massive collected training data and lengthy training time are costly and impractical for realistic deployment. In this paper, a knowledge-driven meta-learning solution for CSI feedback is proposed, where the DL model initialized by the meta model obtained from meta training phase is able to achieve rapid convergence when facing a new scenario during the target retraining phase. Specifically, instead of training with massive data collected from various scenarios, the meta task environment is constructed based on the intrinsic knowledge of spatial-frequency characteristics of CSI for meta training. Moreover, the target task dataset is also augmented by exploiting the knowledge of statistical characteristics of channel, so that the DL model initialized by meta training can rapidly fit into a new target scenario with higher performance using only a few actually collected data in the target retraining phase. The method greatly reduces the demand for the number of actual collected data, as well as the cost of training time for realistic deployment. Simulation results demonstrate the superiority of the proposed approach from the perspective of feedback performance and convergence speed.

show abstract

Section: Introductionmentioning

confidence: 99%

A Knowledge-Driven Meta-Learning Method for CSI Feedback

Han¹,

Tian²,

Li³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Among various deployment problems, unfavorable generalization is one of the significant challenge in DL-based CSI feedback. Online transfer learning [17], [18] strategies are introduced to DL-based CSI feedback to mitigate the influence of channel mismatch. Other deployment problems are also widely studied including joint design with other modules [19]- [21], bit-stream generation [22], multiple-rate feedback [23], imperfect feedback [24], etc.. Beside the aforementioned research, the hardware limitations in computation and storage are also considered to facilitate the deployment of DL-based CSI feedback in practical communication systems.…”

Section: Introductionmentioning

confidence: 99%

Lightweight Neural Network with Knowledge Distillation for CSI Feedback

Cui¹,

Guo²,

Chen³

et al. 2022

Preprint

View full text Add to dashboard Cite

Deep learning-based autoencoder has shown considerable potential in channel state information (CSI) feedback. However, the excellent feedback performance achieved by autoencoder is at the expense of a high computational complexity. In this paper, a knowledge distillation-based neural network lightweight strategy is introduced to deep learning-based CSI feedback to reduce the computational requirement. The key idea is to transfer the dark knowledge learned by a complicated teacher network to a lightweight student network, thereby improving the performance of the student network. First, an autoencoder distillation method is proposed by forcing the student autoencoder to mimic the output of the teacher autoencoder. Then, given the more limited computational power at the user equipment, an encoder distillation method is proposed where distillation is only performed to student encoder at the user equipment and the teacher decoder is directly used at the base stataion. The numerical simulation results show that the performance of student autoencoder can be considerably improved after knowledge distillation and encoder distillation can further improve the feedback performance and reduce the complexity.

show abstract