Rong Peng's research while affiliated with Guangxi Normal University and other places

... With the aid of some recent works [26][27] [28][29] [30], we also want to lay the groundwork for how researchers have implemented deep learning models in the network embedding technique: ...

... The majority of conventional methods are structure embedding models, which can be classified into three categories [26]: translational distance models, semantic matching models, and neural network models. Translational distance models develop distance-based score functions. ...

... Further, in [27] neuro-symbolic approaches dealing with graph structures are classified into three categories. First, logically informed embedding approaches [28] [29] use predefined logical rules that provide knowledge to a neural system, while both components are mostly distinct. Second, approaches for knowledge graph embedding with logical constraints [30] [31] use prior knowledge as constraints on the neural knowledge graph embedding method in order to modify predictions or embeddings. ...

... To comprehensively evaluate our approach, we compare TTEA with Trans-based, GNNs-based and Semi-supervised entity alignment methods: (Huang et al., 2022b), MRGA (Ding et al., 2021), SHEA (Yan et al., 2021), RAGA-l (Zhu et al., 2021a). ...

... The current research on the secure matching of text strings includes algorithmic improvement of secure text matching (STM) [21,22], approximate matching based on Bloom Filter [23][24][25][26], exact matching [27][28][29][30][31][32][33][34][35][36], and string equality problems [37][38][39]. In STM, keyword matching is a very common and important field. ...

... For instance, for each of these datasets, Table IV provides a link to the open-source datasets, along with other meta-data details associated with it, such as primary studies that used it, trace space (the maximum counts of trace links), and other characteristics of the dataset. • P Primary identifier keywords are converted to comment keywords by their similarity in appearance in the syntax tree location Verb-object Phrases [7] • P Extracting verb-object phrases as main information and essential meaning Analyzing Close Relations [13] • G Calculating the close relations (semantic similarity) between target artifacts Term Classification [17] [30] • G Categorizing class names, comments, and all other terms in code Model-Driven Engineering (MDE) [19] • G Combining use of both MDE and IR, analyzing the textual information (organization and hierarchy) contained in the model to retrieve implicit links between documents Hybrid Method [21] [29] • • G Combing VSM and BTM which can help relieve data sparsity caused by short text Genetic Algorithm [29] • G Configuring initial parameters of BTM by introducing Genetic Algorithm Code Calling Relationships [20] • G Identifying errors between requirements and code traces by code-calling relationships Historical Co-change Information [23] • • G Taking the processed corpora and co-change information of classes as input to reorder and filter baseline links Dynamic Integration of Structural and Co-change Coupling [28] • G Retrieving indirect links based on weighted integration of structural coupling and class coupling based on change history Configuration Management Log [35] [38] • • G Restoring links by finding revisions in the configuration management log that contain words related to requirements Frugal User Feedback with Closeness Analysis on Code [40] • ...

... Knowledge integration is based on the thought of the translation methods [25,35,37], using multiple plausibilities to achieve the integration of heterogeneous knowledge. Knowledge alignment originates from the idea of entity alignment [39]. The process of the knowledge reconstruction is shown in Fig. 3. ...

... Comparing with CEA, PCEA could further decrease the execution time and main memory consumption of the tuning process, without sacrificing the quality of alignment. Lv et al. [23] proposed a new meta-matching technology for ontology alignment with grasshopper optimization (GSOOM), which used The Grasshopper Optimization Algorithm (GOA) to find the corresponding relationship between the source ontology and target ontology by optimizing the weight of multiple similarity measures. They modeled the ontology meta-matching problem as an optimization GOA individual fitness problem with two objective functions. ...

... To set a conceptualized framework for context modeling we begin with setting the requirements. In doing so, we refer to existing studies, such as [10,[12][13][14] and to the experience we gained in context-aware systems. In the following, we list the requirements we gathered. ...

... The question answering approach proposed in this study simplifies the task pipeline by converting the genealogical knowledge graph into text, which is then combined with unstructured genealogical texts and processed by BERT's contextual embeddings. Converting the genealogical graph into text passages can be performed using knowledge-graph-to-text templates and methodologies [21,26,55,76,123], and knowledge-graph-to-text machine learning and DNN models [5,33,63,66,68,78,79,99,106]. Templatebased knowledge-graph-to-text methods use hardcoded or extracted linguistic rules or templates to convert a subgraph into a sentence. ...