Knowledge Graph

A knowledge graph is a structured representation of facts, consisting of entities, relationships, and semantic descriptions.

• Property graphs or attributed graphs are widely used, in which nodes and relations have properties or attributes.
• Knowledge can be expressed in a factual triple in the form of (head, relation, tail) or (subject, predicate, object)under the resource description framework (RDF), for example,(Albert Einstein, WinnerOf, Nobel Prize).
• It can also be represented as a directed graph with nodes as entities and edges as relations.

Overview: History, Notation, Definitions, and Categorization

brief history

definition & notations

we define a knowledge graph as $\mathcal{G}={\mathcal{E}, \mathcal{R}, \mathcal{F}}$, where $\mathcal{E}, \mathcal{R}$ and $\mathcal{F}$ are sets of entities, relations and facts, respectively. A fact is denoted as a triple $(h, r, t) \in \mathcal{F}$.

Categorization of Research on Knowledge Graph

1 knowledge representation learning (KRL)

Knowledge Representation Learning (KRL)

Knowledge Representation Learning is a critical research issue of knowledge graph which paves the way for many knowledge acquisition tasks and downstream applications.

We categorize KRL into four aspects: representation space, scoring function, encoding models and auxiliary information, providing a clear workflow for developing a KRL model. Specific ingredients include:

1) representation space: in which the relations and entities are represented;

Representation space includes point-wise space, manifold, complex vector space, Gaussian distribution, and discrete space.

2) scoring function: for measuring the plausibility真实性 of factual triples;

Scoring metrics are generally divided into distance-based and similarity matching based scoring functions.

3) encoding models: for representing and learning relational interactions;

Current research focuses on encoding models, including linear/bilinear models, factorization, and neural networks.

4) auxiliary informatio:n to be incorporated into the embedding methods.

Auxiliary information considers textual, visual, and type information.

2 knowledge acquisition

Knowledge Acquisition tasks are divided into three categories:

1. KGC: knowledge graph completion expanding exisiting knowledge graphs.
   • embedding-based ranking
   • relation path reasoning
   • rule-based reasoning
   • meta relational learning
2. relation extraction discover new knowledge (relations) from the text.
   • Relation extraction models utilize attention mechanism, graph convolutional networks (GCNs), adversarial training, reinforcement learning, deep residual learning, and transfer learning.
3. entity discovery. discover new knowledge (entities) from the text.
   • recognition, disambiguation, typing, and alignment.

3 temporal knowledge graphs

Temporal Knowledge Graphs incorporate temporal information for representation learning. This survey categorizes four research fields, including temporal embedding, entity dynamics, temporal relational dependency, and temporal logical reasoning.

4 knowledge-aware applications

Knowledge-aware Applications include natural language understanding (NLU), question answering, recommendation systems, and miscellaneous real-world tasks, which inject knowledge to improve representation learning.