Downlink channel state information (CSI) is critical in a frequency division duplexing (FDD) massive multiple-input multiple-output (MIMO) system. We exploit the reciprocity between uplink and downlink channels in angular domain and diagnose the supports of downlink channel from the estimated uplink channel. While the basis mismatch effects will damage the sparsity level and the path angle deviations between uplink and downlink transmission paths will induce differences in channel supports, a downlink support diagnosis algorithm based on the DBSCAN (density-based spatial clustering of applications with noise) which is widely used in machine learning is presented. With the diagnosed supports of downlink channel in angular domain, a weighted subspace pursuit (SP) channel estimation algorithm for FDD massive MIMO is proposed. The restricted isometry property (RIP)-based performance analysis for the weighted SP algorithm is given out. Both the analysis and the simulation results show that the proposed downlink channel estimation with diagnosed supports is superior to the standard iteratively reweighted least squares (IRLS) and SP without channel priori or with the assumption of the common supports for uplink and downlink channels in angular domain.
Named entity recognition of military equipment is an important task in the construction of knowledge graph in the military domain. It is a key technical means to improve the intelligence degree of military intelligence information retrieval, intelligence analysis, command and decision. There are many problems in the task of named entity recognition in the field of military equipment, such as fuzzy entity boundary, complex grammar structure and many professional words, which directly lead to the loss of accuracy in named entity recognition. Aiming at the above challenges, this paper proposes a BERT-BILSTM -CRF neural network model based on type labeling and part-of-speech labeling of entities in the field of military equipment. BERT's pre-trained language model fully considers the correlation between words when constructing word vectors, which is used to supplement the semantic relations embedded in words, and can solve the fuzzy problem in name entity recognition. BILSTM layer is used to carry out bidirectional semantic coding, which can solve the longdistance dependence problem. Finally, the output of BERT-BILSTM layer is decoded by CRF layer, and the optimal tag sequence is obtained. The experimental results show that compared with CRF model and BILSTM-CRF model, the F1 value of the proposed model is increased by 10%.Compared with the BILSTM-Attention-CRF model, the F value of this model increased by 10.48%, and the recall rate increased by 15.02%.Compared with the BERT-IDCNN-CRF model, the F value of this model is increased by 0.62%, and the recall rate is increased by 4.55%.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.