Downlink CSI Feedback Algorithm With Deep Transfer Learning for FDD Massive MIMO Systems

Abstract

In this paper, a channel state information (CSI) feedback method is proposed based on deep transfer learning (DTL). The proposed method addresses the problem of high training cost of downlink CSI feedback network in frequency division duplexing (FDD) massive multiple-input multiple-output (MIMO) systems. In particular, we obtain the models of different wireless channel environments at low training cost by finetuning the pre-trained model with a relatively small number of samples. In addition, the effects of different layers on training cost and model performance are discussed. Furthermore, a modelagnostic meta-learning (MAML)-based method is proposed to solve the problem associated with large number of samples of a wireless channel environment required to train a deep neural network (DNN) as a pre-trained model. Our results show that the performance of the DTL-based method is comparable with that of the DNN trained with a large number of samples, which demonstrates the effectiveness and superiority of the proposed method. At the same time, although there is a certain performance loss compared with the DTL-based method, the MAML-based method shows good performance in terms of the normalized mean square error (NMSE).

Overview of Deep Learning-based CSI Feedback in Massive MIMO Systems

“…Online training [59], [138]: transfer learning and meta learning is introduced to accelerate online training at the BS; [139]: the feedback NN is trained at the user side, and FL is adopted; [140]: a new encoder is trained at the user side for a specific area without changing the decoder, and gossip learning applied to multiuser scenario;…”

“…Therefore, online training is essential. Inspired by [151] that applies transfer learning to DL-based CSI prediction, several novel online training strategies are introduced to DL-based CSI feedback in [59] and [138] to accelerate the training convergence. Once the environment changes, pretrained NNs are fine-tuned using new CSI samples, which is regarded as transfer learning.…”

“…Then, the model-agnostic meta learning algorithm, which trains NNs by alternating inner-task and across task updates and then adjusts the original NNs for a new environment with few CSI samples, is adopted to accelerate the NN training further. The training in [59], [138] is implemented at the BS. In [139], the feedback NN is trained at the user side.…”

Machine Learning-Based CSI Feedback With Variable Length in FDD Massive MIMO

“…These works exploit the conjecture that learning at the uplink frequency can be transferred at the downlink frequency with no further modification, as proposed for the first time in [30] and validated in [36]. Furthermore, this problem has been addressed through transfer learning in [37], where the authors investigate the trade-off between training cost and model performance. Second, deep autoencoders compress the channel into a latent space with fixed dimensionality, determined by the number of neurons present in their middle layer.…”

Machine Learning in Beyond 5G/6G Networks—State-of-the-Art and Future Trends

“…To this end, few studies are exploring potential solutions. For example, in [133] a MAML-based method is proposed o solve the challenge of associated large number of samples in a wireless channel environment, in order to train a deep neural network (DNN) with good results in terms of Normalized Mean Squarred Error (NMSE). Furthermore, the authors in [134] propose a new decoder, namely Model Independent Neural Decoder (MIND) based on a MAML methodology achieving satisfactory parameter initialization in the meta-training stage and accuracy results.…”