Prompt paradigm is widely used to zero-shot Natural Language Processing （NLP） tasks. However， the existing zero-shot Relation Extraction （RE） model based on Prompt paradigm suffers from the difficulty of constructing answer space mappings and dependence on manual template selection， which leads to suboptimal performance. To address these issues， a zero-shot RE model via multi-template fusion in Prompt was proposed. Firstly， the zero-shot RE task was defined as the Masked Language Model （MLM） task， where the construction of answer space mapping was abandoned. Instead， the words output by the template were compared with the relation description text in the word embedding space to determine the relation class. Then， the part of speech of the relation description text was introduced as a feature， and the weight between this feature and each template was learned. Finally， this weight was utilized to fuse the results output by multiple templates， thereby reducing the performance loss caused by the manual selection of Prompt templates. Experimental results on FewRel （Few-shot Relation extraction dataset） and TACRED （Text Analysis Conference Relation Extraction Dataset） show that， the proposed model significantly outperforms the current state-of-the-art model， RelationPrompt， in terms of F1 score under different data resource settings， with an increase of 1.48 to 19.84 percentage points and 15.27 to 15.75 percentage points， respectively. These results convincingly demonstrate the effectiveness of the proposed model for zero-shot RE tasks.

Logs can record the specific status of the system during the operation， and automated log anomaly detection is critical to network security. Concerning the problem of low accuracy in anomaly detection caused by the evolution of log sentences over time， an unsupervised log anomaly detection model LogCL was proposed. Firstly， the log parsing technique was used to convert semi-structured log data into structured log templates. Secondly， the sessions and fixed windows were employed to divide log events into log sequences. Thirdly， quantitative characteristics of the log sequences were extracted， natural language processing technique was used to extract semantic features of log templates， and Term Frequency-Inverse Word Frequency （TF-IWF） algorithm was utilized to generate weighted sentence embedding vectors. Finally， the feature vectors were input into a parallel model based on Convolutional Neural Network （CNN） and Bi-directional Long Short-Term Memory （Bi-LSTM） network for detection. Experimental results on two public real datasets show that the proposed model improves the anomaly detection F1-score by 3.6 and 2.3 percentage points respectively compared with the baseline model LogAnomaly. Therefore， LogCL can perform effectively on log anomaly detection.