A Survey on Knowledge Graph Embeddings for Link Prediction

In the paper

“However, there is no overall classification and summary of the KGE models proposed in recent years in this paper, and the selected models are few, which cannot cover all types of KGE models.”

What the sources say

• Disputed

  Section 3 (p. 5) explicitly proposes a taxonomy of KG link-prediction models, dividing them into three main families and smaller groups. The Introduction (p. 2) also describes it as an educational taxonomy for knowledge graph embedding techniques, contrary to the claim that the paper offers no overall classification. The paper selects 16 representative models, but does not say they cover every type. (LLM, conf=0.89)

  Checked by a reasoning model — may be wrong.

  cites [28] Rossi, 2021, Knowledge graph embedding for link prediction: A comparative analysis, ACM Trans. Knowl. Discov. Data TKDD sourceExperimental — accuracy not yet measured

In the paper

“TransW [101] uses word embeddings to compose knowledge graph embeddings and learns a function mapping from the word embedding space to the knowledge embedding space.Entities and relations are represented in the form of linear combinations of word embeddings in this model, which can detect unknown facts”

What the sources say

• Supported

  The Abstract (p. 1) says TransW composes knowledge graph embeddings using word embeddings and learns a mapping between the two spaces. The TransW section (p. 3) describes entities and relations as sums of transformed word embeddings, and Detecting Unknown Facts (p. 4) says this composition makes it possible to detect facts with unseen entities or relations. (LLM, conf=0.94)

  cites [101] Ma, L., Sun, P., Lin, Z., and Wang, H. (2019). Composing Knowledge Graph Embeddings via Word Embeddings. arXiv. source

In the paper

“The traditional semantic-based models focus only on the information of the triple itself and do not fuse any additional information, as DistMult [20] and Complex [21] do.”

What the sources say

• Supported

  Section 3 (p. 3) describes learning entity and relation representations from KB triples, and Section 4.1 (p. 5) identifies the paper’s baseline DISTMULT as the diagonal bilinear model. Section 3.1 (p. 3) says most such models use one-hot entity inputs; the paper treats pre-trained vectors as separate DISTMULT variants, so the claim fits baseline DistMult rather than every variant. (LLM, conf=0.87)

  cites [20] Yang, B., Yih, W., He, X., Gao, J., and Deng, L. (2015, January 7–9). Embedding Entities and Relations for Learning and Inference in Knowledge Bases. Proceedings of the ICLR (Poster), San Diego, CA, USA. source
• Supported

  Section 3 (p. 4) describes ComplEx as learning from observed knowledge-base triples and scoring a candidate using its subject, relation, and object embeddings. The scoring-function table in Section 2.1 (p. 3) shows the same three inputs for both DistMult and ComplEx, with no additional information fused into either scoring function. (LLM, conf=0.86)

  cites [21] Trouillon, T., Welbl, J., Riedel, S., Gaussier, É., and Bouchard, G. (2016). Complex Embeddings for Simple Link Prediction, ICML. source

In the paper

“TransH introduces a hyperplane, and TransR uses a relation-specific space to handle different relations, excavating more semantic information.”

What the sources say

• Partly supported

  The TransH section (p. 3) confirms that the paper introduces relation-specific hyperplanes and projects entity embeddings onto them to represent different roles across relations. The cited paper does not address TransR or its use of a relation-specific space, so it supports only the TransH part of the sentence. (LLM, conf=0.90)

  cites [15] Wang, Z., Zhang, J., Feng, J., and Chen, Z. (2014). Knowledge Graph Embedding by Translating on Hyperplanes, AAAI Press. sourceExperimental — accuracy not yet measured
• Supported

  The TransE and TransH section (p. 2) describes TransH projecting entity embeddings onto relation hyperplanes. The Introduction (p. 1) and TransR section (p. 3) describe TransR using relation-specific spaces to represent different aspects of entities; the Link Prediction section (p. 4) reports more precise representations and improved performance. (LLM, conf=0.93)

  cites [17] Lin, Y., Liu, Z., Sun, M., Liu, Y., and Zhu, X. (2015). Learning Entity and Relation Embeddings for Knowledge Graph Completion, AAAI Press. source

In the paper

“LiteralE [115] introduces the literature information, giving priority to numerical literature.”

What the sources say

• Supported

  The Introduction (p. 2) says LiteralE enriches entity embeddings with literal information and that the paper mainly focuses on numerical literals. In context, the citing sentence’s “literature information” appears to mean literal information. (LLM, conf=0.91)

  cites [115] Kristiadi, A., Khan, M.A., Lukovnikov, D., Lehmann, J., and Fischer, A. (2019, January 26–30). Incorporating Literals into Knowledge Graph Embeddings. Proceedings of the ISWC, Auckland, New Zealand. source

In the paper

“Ji et al. [31] divided the research on KGs into four categories: knowledge representation learning, knowledge acquisition, temporal KGs and knowledge-aware applications.”

What the sources say

• Supported

  The Introduction’s “Full-view categorization and new taxonomies” passage explicitly says the survey reviews knowledge-graph research in four high-level aspects: KRL, knowledge acquisition, temporal knowledge graphs, and knowledge-aware applications. These match the four categories attributed to Ji et al.; no page number is provided in our copy. (LLM, conf=0.95)

  cites [31] Ji, S., Pan, S., Cambria, E., Marttinen, P., and Yu, P.S. (2020). A Survey on Knowledge Graphs: Representation, Acquisition and Applications. arXiv. source

In the paper

“In the encoder part, by introducing weights for different types of relations, SACN improves the GCN model to obtain a weighted graph convolutional network (WGCN).”

What the sources say

• Supported

  The Method section (“Weighted Graph Convolutional Layer”) explicitly calls WGCN an extension of classic GCN that weighs different relation types differently during aggregation. The Introduction identifies WGCN as SACN’s encoder. (LLM, conf=0.95)

  cites [26] Shang, C., Tang, Y., Huang, J., Bi, J., He, X., and Zhou, B. (2019). End-to-End Structure-Aware Convolutional Networks for Knowledge Base Completion, AAAI Press. source

In the paper

“HypER implements nonlinear (quadratic) combinations of entity and relation embeddings rather than the linear combinations (weighted sums) in ConvE, which gives HypER the advantages of a much richer expressive capacity and fewer parameters.”

What the sources say

• Supported

  The Related Work section (p. 3) explicitly says HypER combines entity and relation embeddings nonlinearly (quadratically), unlike ConvE’s linear weighted sum, giving HypER more expressive power while reducing parameters. (LLM, conf=0.95)

  cites [24] Balazevic, I., Allen, C., and Hospedales, T.M. (2019, January 17–19). Hypernetwork Knowledge Graph Embeddings. Proceedings of the ICANN (Workshop), Munich, Germany. source

In the paper

“Rossi et al. [28] classified models into three categories: tensor decomposition models, geometric models and deep learning models.”

What the sources say

• Supported

  The Overview of Link Prediction Techniques (p. 5) explicitly identifies three main families: Tensor Decomposition Models, Geometric Models, and Deep Learning Models. Sections 3.1–3.3 then discuss each family separately. (LLM, conf=0.95)

  cites [28] Rossi, 2021, Knowledge graph embedding for link prediction: A comparative analysis, ACM Trans. Knowl. Discov. Data TKDD source

In the paper

“The second group of multiplicative models includes DistMult [20] and Complex [21], which can outperform the additive models by capturing more semantic information [19].”

What the sources say

• ?Not checked — no full text of the sourcecites [19] Liu, Z., Sun, M., Lin, Y., and Xie, R. (2016). Knowledge Representation Learning: A Review. J. Comp. Res. Develop., 247–261. source
• Partly supported

  The Results (p. 5) report that DistMult outperformed the additive DistAdd model on the evaluated datasets. The Introduction (p. 1) says bilinear embeddings capture the compositional semantics of relations, but the paper does not evaluate Complex or establish that both named models outperform additive models by capturing more semantic information. (LLM, conf=0.86)

  cites [20] Yang, B., Yih, W., He, X., Gao, J., and Deng, L. (2015, January 7–9). Embedding Entities and Relations for Learning and Inference in Knowledge Bases. Proceedings of the ICLR (Poster), San Diego, CA, USA. sourceExperimental — accuracy not yet measured
• Supported

  The citation to this paper is attached to the identification of ComplEx, not to the later explanation of why multiplicative models may outperform additive ones, which cites [19]. Section 2.1 (p. 2) explicitly identifies ComplEx as the paper’s model, and Section 3 (p. 4) describes its embedding-based scoring function and relationship to DistMult. (LLM, conf=0.93)

  cites [21] Trouillon, T., Welbl, J., Riedel, S., Gaussier, É., and Bouchard, G. (2016). Complex Embeddings for Simple Link Prediction, ICML. source

In the paper

“ProjE [176] complements the missing information in a KG by learning the joint embedding of the entities and edges and making modifications to the loss function, thus improving the KGE model through trivial changes to the network architecture and eliminating complex feature engineering.”

What the sources say

• Supported

  Split into 2 atoms — 2 confirmed, 0 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  cites [176] Shi, B., and Weningr, T. (2017). ProjE: Embedding Projection for Knowledge Graph Completion, AAAI Press. source

In the paper

“ConvE obtains local relationships in different dimensions between entities through a convolution layer and a fully connected layer, which ignores the global relationships between triple embeddings of the same dimension.”

What the sources say

• Partly supported

  The model description (p. 3) says ConvE uses convolutional and fully connected layers to model interactions between entity and relation embeddings. The discussion of 2D convolutions (p. 2) describes interactions between embeddings, but does not say ConvE ignores global relationships between triple embeddings of the same dimension. The Conclusion (p. 7) proposes increasing large-scale structure as future work, which does not establish that specific limitation. (LLM, conf=0.87)

  cites [22] Dettmers, T., Minervini, P., Stenetorp, P., and Riedel, S. (2017). Convolutional 2D Knowledge Graph Embeddings, AAAI Press. sourceExperimental — accuracy not yet measured

In the paper

“The decoder, called Conv-TransE, removes the reshaping operation from the input entity and relation embeddings and lets the convolutional filters operate directly on input entities and relations in the same dimension.Thus, the translation properties of TransE remain, while the same prediction performance as ConvE is maintained.”

What the sources say

• Supported

  Split into 2 atoms — 2 confirmed, 0 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  cites [26] Shang, C., Tang, Y., Huang, J., Bi, J., He, X., and Zhou, B. (2019). End-to-End Structure-Aware Convolutional Networks for Knowledge Base Completion, AAAI Press. source

In the paper

“Weighted GCN, known as SACN [26], is an end-to-end structure-aware convolutional network that takes into account node connectivity, node attributes and relation types simultaneously.”

What the sources say

• Partly supported

  The Introduction and Conclusion describe SACN as an end-to-end network whose weighted GCN encoder uses graph connectivity, node attributes, and relation types. But the citing sentence conflates the encoder with the full model: WGCN is a component of SACN, which also has a Conv-TransE decoder. (LLM, conf=0.88)

  cites [26] Shang, C., Tang, Y., Huang, J., Bi, J., He, X., and Zhou, B. (2019). End-to-End Structure-Aware Convolutional Networks for Knowledge Base Completion, AAAI Press. sourceExperimental — accuracy not yet measured

In the paper

“Removing the reshaping operation from ConvE, ConvKB [23] uses 1D convolution to maintain the translation characteristics of TransE, which is sufficient for capturing the global relationships and transitional characteristics between entities.”

What the sources say

• Supported

  The Proposed ConvKB model section describes 1 × 3 filters applied across the embedding triple to examine global relationships and generalize the transitional characteristics of transition-based models. It also states that ConvKB can reduce to TransE under specified settings. The Introduction contrasts ConvE’s reshaping operation with ConvKB’s k × 3 input matrix; page numbers are not provided in our copy. (LLM, conf=0.90)

  cites [23] Nguyen, D.Q., Nguyen, T.D., Nguyen, D.Q., and Phung, D.Q. (2018, January 1–6). A Novel Embedding Model for Knowledge Base Completion Based on Convolutional Neural Network. Proceedings of the NAACL-HLT, New Orleans, LA, USA. source

In the paper

“Moreover, ConvE introduces a 1-N scoring program that takes a head-entity-relation pair and matches all the tail entities simultaneously, which is different from other models that use the 1-1 scoring program; thus, much evaluation time is saved.”

What the sources say

• Supported

  The “Fast Evaluation for Link Prediction Tasks” section (p. 3) says ConvE scores one subject–relation pair against all candidate entities simultaneously, rather than scoring triples one at a time. It reports evaluation time falling from 3.34 hours to 35 seconds with 1-N scoring, supporting the claimed time saving. (LLM, conf=0.94)

  cites [22] Dettmers, T., Minervini, P., Stenetorp, P., and Riedel, S. (2017). Convolutional 2D Knowledge Graph Embeddings, AAAI Press. source

In the paper

“For link-prediction tasks, a relational graph convolutional network (R-GCN) can be regarded as an autoencoder consisting of an encoder and a decoder.”

What the sources say

• Supported

  The introductory passage (p. 1) explicitly describes the paper’s link-prediction model as an autoencoder with an R-GCN encoder and a tensor-factorization decoder. Figure 3b (p. 3) likewise depicts an R-GCN encoder paired with a DistMult decoder. (LLM, conf=0.95)

  cites [158] Schlichtkrull, M.S., Kipf, T.N., Bloem, P., Berg, R.v.d., Titov, I., and Welling, M. (2018, January 3–7). Modeling Relational Data with Graph Convolutional Networks. Proceedings of the ESWC, Crete, Greece. source

In the paper

“In addition, the experiment of this survey was not reproduced in a unified environment configuration but used different coding frameworks, such as Python and C++.”

What the sources say

• Partly supported

  Section 5.1 (pp. 18–19) says Rossi et al. trained and evaluated the models from scratch using implementations from several repositories, which supports the use of different implementations. But it says all experiments ran on the same server environment; it does not establish that the environment configuration was not unified or that the implementations used Python and C++. (LLM, conf=0.83)

  cites [28] Rossi, 2021, Knowledge graph embedding for link prediction: A comparative analysis, ACM Trans. Knowl. Discov. Data TKDD sourceExperimental — accuracy not yet measured

In the paper

“SimplEx [152] has low performance for FB15k-237 and high performance for FB15k because FB15k contains inverse relations and SimplEx can model the inverse relations appropriately.”

What the sources say

• Partly supported

  Section 6.1 (p. 8) reports strong SimplE performance on FB15k, and Section 4 (p. 3) says the model uses inverse relations. The paper reports experiments on WN18 and FB15k, not FB15k-237 (Section 6, p. 7); it does not establish low performance on FB15k-237 or attribute its FB15k results to inverse relations in that dataset. (LLM, conf=0.88)

  cites [152] Kazemi, S.M., and Poole, D. (2018). SimplE Embedding for Link Prediction in Knowledge Graphs. NeurIPS. sourceExperimental — accuracy not yet measured

In the paper

“SACN [26] also performs well among these models; it uses the relation types and entity attributes in the GCN model structure.”

What the sources say

• Supported

  The Introduction describes SACN’s weighted graph convolutional encoder as using node attributes and relation types. The Experiments > Results section reports that SACN performed best against the baseline models tested, supporting the statement that it performs well. (LLM, conf=0.92)

  cites [26] Shang, C., Tang, Y., Huang, J., Bi, J., He, X., and Zhou, B. (2019). End-to-End Structure-Aware Convolutional Networks for Knowledge Base Completion, AAAI Press. source

In the paper

“The bilinear+TR model has the shortest training time, with a type regularizer incorporated into the loss function, which fully considers the type information of entities.”

In the paper

“Semantic matching models such as DistMult and ComplEx all suffer from longer training times.”

In the paper

“Translational distance models such as HAKE and RotatE all have shorter training times because the translational-distance model has a relatively simple model structure and scoring function without too many parameters.”

In the paper

“LiteralE introduces some overhead in terms of the number of parameters compared to the base method, leading to a long training time.”

In the paper

“The performance of conventional models such as translation models (TransE) and semantic models (DistMult and ComplEx) is not good.”

In the paper

“While HypER utilizes a 1D relation-specific filter and a nonlinear (quadratic) combination of entity and relation embeddings via hypernetworks to perform weight sharing, it has many fewer parameters than ConvE, so it saves much training time.”

In the paper

“For FB15k, only ComplEx obtains good results upon adding literal information, while for FB15-237, only DistMult improves slightly after adding literal information.”

In the paper

“Since the traditional model was put forward, semantic-matching-based models have been continuously mining deeper semantic information, including path information, entity types, contextual information and so on, to obtain more accurate representations for KGE.”

In the paper

“HAKE [60] considers the semantic hierarchy of entities by using polar coordinates, in which concentric circles can naturally reflect the hierarchy.”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites [60] Zhang, Z., Cai, J., Zhang, Y., and Wang, J. (2020, January 7–12). Learning Hierarchy-Aware Knowledge Graph Embeddings for Link Prediction. Proceedings of the AAAI 2020, New York, NY, USA.

In the paper

“The former recently proposed models fuse various additional information to obtain better performance to mine deeper semantic information at the bottoms of graphs.”

In the paper

“These models are additive models, which usually use distance-based functions to define the scoring function for link prediction.”

In the paper

“RW-LMLM is a novel approach for link prediction that consists of a random walk algorithm for KGs (RW) and a language model-based link-prediction model (LMLM).”

• Also stated in: § 3.2. Semantic Information-Based Models · plain

  “RW can capture the semantic and syntactic information in KGs by considering the entities, relations and order information of the paths, in contrast to DeepWalk (which considers only the entities and relations).”

In the paper

“Gaussian embedding models take the uncertainties of the entities and relations into account, but this results in a complex model.”

In the paper

“Second, real-world data are often dynamic and evolving, which leads to difficulty in constructing correct and complete KGs [7].”

What the sources say

• ?Not checked — no full text of the sourcecites [7] Cai, 2017, A Comprehensive Survey of Graph Embedding: Problems, Techniques, and Applications, IEEE Trans. Knowl. Data Eng. source

In the paper

“In this way, it considers the semantics between the head entity and the relation as well as the semantics between the relation and the tail entity, which is not done in previous models.”

What the sources say

• ?Not checked — no full text of the sourcecites [52] Yang, S., Tian, J., Zhang, H., Yan, J., He, H., and Jin, Y. (2019, January 10–16). TransMS: Knowledge Graph Embedding for Complex Relations by Multidirectional Semantics. Proceedings of the IJCAI, Macao, China. source

In the paper

“Semantic information-based models usually use similarity-based functions to define scoring functions for traditional semantic-matching models or introduce additional information to mine more knowledge for recently proposed models.”

In the paper

“TransE [13] is one of the earliest models for link prediction, which uses the relation for translating the head entity to a tail entity.”

What the sources say

• Not checked yetcites [13] Bordes, A., Usunier, N., Garcia-Duran, A., Weston, J., and Yakhnenko, O. (2013, January 5–8). Translating Embeddings for Modeling Multi-Relational Data. Proceedings of the NIPS, Lake Tahoe, NV, USA.

In the paper

“These models suffer from higher computational complexity.”

In the paper

“The TransC [149] model combines structural information with entity concepts (referring to the categories of entities) to improve KGE models and introduces a novel type of semantic similarity to measure the distinctness of entity semantics by using the concept information.”

What the sources say

• ?Not checked — no full text of the sourcecites [149] Wang, Y., Liu, Y., Zhang, H., and Xie, H. (2019). Leveraging Lexical Semantic Information for Learning Concept-Based Multiple Embedding Representations for Knowledge Graph Completion. Asia-Pacific Web (APWeb) and Web-Age Information Management (WAIM) Joint International Conference on Web and Big Data, Springer. source

In the paper

“Although the embedding operation of this type of model is relatively simple, the expressiveness is not sufficient due to the direct addition of h + r and the lack of different semantic information.”

In the paper

“TransE is the earliest translation-based embedding model, and it has difficulty dealing with multirelational graphs; it is limited by its simple translation operation as well as its lack of a discrimination policy for all kinds of relations.”

In the paper

“These three categories reflect the mainstream models of KGE development in three main stages.”

In the paper

“These models consider the type of entity or relation, temporal information, path information and substructure information.”

In the paper

“Therefore, these datasets could be used to evaluate the performance of the model more comprehensively.”

In the paper

“The additional information includes path information, order information, concepts, entity attributes, entity types and so on.”

• Also stated in: § 3.3. Neural Network-Based Models · plain

  “The additional information also includes path information, order information, concepts and so on.”

In the paper

“We also used the code provided in the original paper with the same settings to obtain the best results.”

In the paper

“There are few studies on dynamic graphs, which are better able to reflect the real world.”

In the paper

“Regarding additional information, they considered only these two aspects and also did not provide an overall table.”

What the sources say

• ?Not checked — no full text of the sourcecites [19] Liu, Z., Sun, M., Lin, Y., and Xie, R. (2016). Knowledge Representation Learning: A Review. J. Comp. Res. Develop., 247–261. source

In the paper

“It can be scaled linearly on the edge of the graph to learn hidden layer representations for encoding the local graphical structures and features of nodes.”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Kipf et al.

In the paper

“KGE models embed entities and relations into a low-dimensional vector space while preserving the structure of the KG and its underlying semantic information.”

In the paper

“Dai et al. [19] described these models with two categories: triplet fact-based representation learning models and description-based representation learning models.”

What the sources say

• ?Not checked — no full text of the sourcecites [19] Liu, Z., Sun, M., Lin, Y., and Xie, R. (2016). Knowledge Representation Learning: A Review. J. Comp. Res. Develop., 247–261. source

In the paper

“It can model N-1, N-N, 1-N-1 and one-relation circle structures with the advantage of a combination operator design.”

What the sources say

• ?Not checked — no full text of the sourcecites [178] Zhang, W., Li, J., and Chen, H. (2018, January 26–30). ProjR: Embedding Structure Diversity for Knowledge Graph Completion. Proceedings of the NLPCC, Hohhot, China. source

In the paper

“Traditional distance-based and semantic-matching-based models cannot meet the requirements of KGE.”

In the paper

“The use of convolutional neural networks or attention mechanisms also helps to generate better embeddings.”

In the paper

“Link prediction is one of the most common tasks for evaluating the performance of KGE, which has no formal definition.”

In the paper

“It is difficult to model the heterogeneity of graphs.”

In the paper

“Neural networks have been a promising solution in many fields and have the abilities of self-study, associative storage and high-speed optimization.”

In the paper

“In addition, our experiments on some representative models were conducted in a unified environment, including the server type and programming language (pytorch).”

In the paper

“Recently, to obtain better and more effective entity and relation embeddings, a neural-network model was also introduced into KGE to propagate neighborhood information.”

In the paper

“We conducted all the experiments targeting the link-prediction task in a unified environment with a T640 Dell server, the Ubuntu 16.04 system, and an NVIDIA-SMI 418.67 GPU.”

In the paper

“As described above, there are many models for KGE, but there are still many challenges in obtaining better entity and relation embeddings for link prediction.”

In the paper

“Kipf et al. proposed a simple but effective hierarchical-propagation rule running directly on a graph with a neural-network model, a GCN.”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Kipf et al.

In the paper

“Especially, the graph convolution neural network has a strong ability to mine the underlying semantics of knowledge graphs.”

In the paper

“In addition, if the node information of a multi-hop domain can be aggregated, the accuracy of the model in specific tasks can be greatly improved.”

In the paper

“Therefore, these selected models are representative for gaining deep insight into the existing models, and they are conventional, recently proposed or highly cited.”

In the paper

“which indicates that the rotation operation used in translational-distance models and the hypernetwork used in CNN-based models play important roles in improving performance.”

In the paper

“RotatE uses the complex space and mines different types of relations (symmetry/antisymmetry, inversion and composition); thus, different aspects of semantic information are modeled well by this integration strategy.”

In the paper

“Its scoring function can only capture pairwise interactions of the same dimension between entities.Therefore, this simple embedding can only deal with symmetric relations.”

In the paper

“Compared with traditional models, the latest models have their own advantages.Generally, they have achieved better results, benefiting from their own unique model structures and sampling technologies or adding important additional information.”

In the paper

“Because ComplEx introduces a complex vector space and can deal with asymmetric relations, it has a good response to literal information for FB15k but not FB15k-237.”

In the paper

“Models with additional information, such as node attributes, node types, relationship types, prior knowledge and so on, have better performance.”

In the paper

“RW-LMLM [91] takes into account both the order information and random walk algorithm, and it has the capability of dealing with underlying semantic information.”

What the sources say

• ?Not checked — no full text of the sourcecites [91] Wang, C., Yan, M., Yi, C., and Sha, Y. (2019, January 26–30). Capturing Semantic and Syntactic Information for Link Prediction in Knowledge Graphs. Proceedings of the ISWC, Auckland, New Zealand. source

In the paper

“For HAKE [60], we believe that the polar coordinates may have great benefits because of their particular structure, which enables them to mine considerable hidden semantic structure information.”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites [60] Zhang, Z., Cai, J., Zhang, Y., and Wang, J. (2020, January 7–12). Learning Hierarchy-Aware Knowledge Graph Embeddings for Link Prediction. Proceedings of the AAAI 2020, New York, NY, USA.