International Curriculum (July 2026)

Knowledge Graphs, Entities, and Semantic AI Engineering

Level: Beginner → World-Class Research Engineer
Duration: 12 months (4 quarters)
Outcome: Design, build, and deploy industrial knowledge graphs for AI agents, search engines, robotics, science, medicine, and AGI.

Semester 1 — Foundations

Module 1. Information as Reality

• What information is

• Objects vs concepts

• Identity

• Attributes

• Relations

• Events

• Time

• Space

• Context

• Abstraction

• Hierarchies

Project:

• Describe the real world without databases.

Module 2. Entities

Study

• Entity

• Canonical entity

• Alias

• Identifier

• URI

• UUID

• Entity resolution

• Entity linking

• Entity disambiguation

• Entity extraction

• Named Entity Recognition (NER)

Practice

Create an entity catalog of

• people

• companies

• products

• cities

• scientific concepts

Module 3. Relations

Types

• is-a

• part-of

• instance-of

• causes

• depends-on

• located-in

• created-by

• owned-by

• influences

• similar-to

• opposite-of

Project

Build a graph of 50 relation types.

Module 4. Ontologies

Learn

• ontology

• taxonomy

• vocabulary

• schema

• classes

• subclasses

• inheritance

• restrictions

• reasoning

Standards

• RDF

• RDFS

• OWL

• SHACL

• SKOS

Semester 2 — Knowledge Graph Engineering

Module 5. Graph Theory

Topics

• vertices

• edges

• paths

• cycles

• centrality

• clustering

• communities

• graph traversal

• shortest paths

Algorithms

• BFS

• DFS

• PageRank

• Connected Components

• Louvain

• HITS

Module 6. Graph Databases

Master

• Neo4j

• Memgraph

• Amazon Neptune

• ArangoDB

• TigerGraph

• JanusGraph

Languages

• Cypher

• Gremlin

• SPARQL

Projects

Create

• Movie Graph

• Company Graph

• Medical Graph

Module 7. Knowledge Representation

Represent

• facts

• uncertainty

• confidence

• provenance

• evidence

• sources

• trust

Learn

• RDF triples

• quads

• hypergraphs

• property graphs

Module 8. Data Integration

Topics

• ETL

• ELT

• schema matching

• entity matching

• record linkage

• deduplication

• canonicalization

Practice

Merge five public datasets into one knowledge graph.

Semester 3 — AI + Knowledge Graphs

Module 9. Knowledge Graph Embeddings

Study

• TransE

• RotatE

• DistMult

• ComplEx

• Node2Vec

• DeepWalk

• GraphSAGE

Project

Train embeddings.

Module 10. Graph Neural Networks

Topics

• GCN

• GAT

• Graph Transformers

• Message Passing

• Heterogeneous Graphs

Frameworks

• PyTorch Geometric

• DGL

Module 11. LLM + Knowledge Graphs

Learn

• RAG

• GraphRAG

• Hybrid Search

• Semantic Search

• Agent Memory

• Tool Calling

• Knowledge Retrieval

Projects

Build

• AI assistant

• GraphRAG chatbot

• Enterprise search

Module 12. Vector Databases

Study

• embeddings

• cosine similarity

• HNSW

• IVF

• PQ

• ANN search

Systems

• Qdrant

• Milvus

• Pinecone

• Weaviate

• Chroma

Semester 4 — Advanced Research

Module 13. Internet-scale Knowledge Graphs

Study

• Google Knowledge Graph

• Wikidata

• DBpedia

• OpenAlex

• OpenStreetMap

• Schema.org

Module 14. Semantic Search

Topics

• BM25

• Dense Retrieval

• Hybrid Retrieval

• Cross Encoder

• Re-ranking

Module 15. AI Agents

Build

• autonomous agents

• planning

• memory

• reasoning

• graph memory

• long-term memory

• multi-agent systems

Module 16. Future Knowledge Systems

Research

• World Models

• Knowledge Engines

• Scientific Knowledge Graphs

• Autonomous Discovery

• Machine Scientists

• AGI Memory Architectures

Programming Stack

• Python

• SQL

• Cypher

• SPARQL

• RDF

• Turtle

• JSON-LD

• YAML

• GraphQL

Libraries

• NetworkX

• Neo4j Python Driver

• rdflib

• spaCy

• Hugging Face Transformers

• PyTorch

• PyTorch Geometric

• LangChain

• LlamaIndex

• Haystack

• DSPy

Infrastructure

• Docker

• Kubernetes

• Linux

• Git

• GitHub

• APIs

• Apache Kafka

• Apache Spark

• Airflow

Mathematics

• Linear Algebra

• Graph Theory

• Probability

• Information Theory

• Statistics

• Optimization

• Machine Learning

• Spectral Graph Theory

Final Capstone

Design and implement a production-scale Knowledge Operating System (Knowledge OS) that includes:

• ontology engineering

• entity extraction pipeline

• entity resolution

• relation extraction

• graph database

• vector database

• GraphRAG

• AI agents

• semantic search

• reasoning engine

• API

• visualization dashboard

• continuous graph updates

• benchmarking

• documentation

• deployment in the cloud

Competencies after graduation

By the end of this program, you will be able to:

• Design enterprise-scale ontologies.

• Build knowledge graphs containing millions or billions of entities.

• Integrate structured, semi-structured, and unstructured data into a unified graph.

• Develop AI systems that combine large language models with symbolic knowledge.

• Build GraphRAG architectures and long-term memory for AI agents.

• Apply graph embeddings and graph neural networks to prediction and reasoning tasks.

• Engineer semantic search systems that outperform keyword-only search.

• Conduct original research in knowledge representation, semantic AI, and next-generation intelligent systems.

glossary 

1. Information

Definition

Information is a structured description that reduces uncertainty about the state, properties, relationships, or behavior of one or more entities.

Parent

• Representation

Has

• meaning

• structure

• context

• source

Related

• Data

• Knowledge

• Signal

2. Data

Definition

Data are recorded observations, measurements, symbols, or values that represent aspects of reality without necessarily conveying interpreted meaning.

Parent

• Representation

Examples

• 25

• "London"

• 101001

• Image pixels

• Audio samples

3. Knowledge

Definition

Knowledge is organized information connected through validated relationships that enables understanding, reasoning, prediction, or decision-making.

Knowledge = Information + Structure + Relations + Context

4. Wisdom

Definition

Wisdom is the ability to apply knowledge appropriately to achieve goals under changing conditions while considering long-term consequences.

5. Meaning (Semantics)

Definition

Meaning is the interpretation that connects a representation to the concept, object, event, or relationship it denotes.

6. Concept

Definition

A concept is an abstract mental or computational category representing a class of entities, properties, events, or relationships sharing defining characteristics.

Examples

Vehicle

Planet

Disease

Justice

Algorithm

7. Entity

Definition

An entity is a uniquely identifiable object, concept, event, process, place, person, organization, or phenomenon that can possess properties and participate in relationships.

Entity = Identity + Properties + Relationships

8. Object

Definition

An object is a concrete or abstract entity considered as a single identifiable unit.

9. Instance

Definition

An instance is a specific realization of a concept or class.

Example

Class

Car

Instance

Tesla Model S VIN XXXXX

10. Class

Definition

A class is a category whose members share defining characteristics and constraints.

11. Category

Definition

A category is a grouping of entities based on shared characteristics.

12. Type

Definition

A type is a formal specification describing the permissible characteristics and behavior of entities.

13. Attribute (Property)

Definition

An attribute is a characteristic or feature describing an entity.

Examples

height

age

weight

color

temperature

14. Value

Definition

A value is the specific content assigned to an attribute.

Example

Height → 183 cm

15. Identifier

Definition

An identifier is a unique symbol or sequence used to distinguish one entity from all others within a given scope.

Examples

UUID

URI

ISBN

DOI

VIN

16. Relationship

Definition

A relationship is a semantic connection linking two or more entities.

Examples

owns

created

located_in

part_of

friend_of

17. Triple

Definition

A triple is the smallest semantic statement consisting of:

Subject → Predicate → Object

Example

Tesla → manufactures → Model S

18. Graph

Definition

A graph is a mathematical structure consisting of nodes connected by edges representing relationships.

19. Knowledge Graph

Definition

A knowledge graph is a graph-based representation of entities, their attributes, relationships, and semantics that supports querying, reasoning, integration, and intelligent inference.

Knowledge Graph =
Entities

• Relationships

• Semantics

• Schema

• Identifiers

20. Node

Definition

A node is the graph representation of an entity.

21. Edge

Definition

An edge is a graph representation of a relationship connecting nodes.

22. Ontology

Definition

An ontology is a formal, explicit specification of concepts, relationships, constraints, and rules describing a domain of knowledge.

Purpose

Shared understanding.

Machine reasoning.

Interoperability.

23. Taxonomy

Definition

A taxonomy is a hierarchical classification system organizing entities into parent–child categories.

24. Schema

Definition

A schema is a formal structure defining the organization, constraints, and allowed relationships within a dataset or knowledge graph.

25. Vocabulary

Definition

A vocabulary is a standardized collection of terms and their meanings used consistently within a domain.

26. Semantics

Definition

Semantics is the study and formal representation of meaning and the relationships between symbols and what they denote.

27. Syntax

Definition

Syntax is the set of rules governing the valid arrangement of symbols independently of their meaning.

28. Context

Definition

Context is the set of surrounding conditions, assumptions, relationships, and circumstances required to correctly interpret information.

29. Identity

Definition

Identity is the persistent uniqueness that allows an entity to be distinguished from all other entities across time and contexts.

30. Entity Resolution

Definition

Entity resolution is the process of determining whether different records refer to the same real-world entity.

31. Entity Linking

Definition

Entity linking is the process of connecting mentions in data or text to canonical entities in a knowledge base.

32. Named Entity Recognition (NER)

Definition

Named Entity Recognition is the task of detecting and classifying mentions of entities in unstructured text into predefined categories such as people, organizations, locations, products, or dates.

33. Embedding

Definition

An embedding is a numerical vector representation of an entity, document, image, or other object that preserves semantic relationships in a continuous vector space.

34. Vector

Definition

A vector is an ordered sequence of numerical values representing a point, direction, or feature set in a mathematical space.

35. Vector Space

Definition

A vector space is a mathematical space in which objects are represented as vectors, enabling operations such as measuring similarity, distance, and transformation.

36. Similarity

Definition

Similarity is a quantitative measure of how closely two entities resemble each other according to a defined metric.

37. Semantic Search

Definition

Semantic search is an information retrieval approach that ranks results according to their meaning and contextual relevance rather than exact keyword matching.

38. Reasoning (Inference)

Definition

Reasoning is the process of deriving new conclusions, relationships, or facts from existing knowledge using logical, statistical, or computational methods.

39. Knowledge Base

Definition

A knowledge base is an organized repository of structured and/or unstructured knowledge designed to support storage, retrieval, reasoning, and decision-making.

40. Artificial Intelligence Knowledge

Definition

Artificial intelligence knowledge is information encoded in computational forms—such as parameters, vectors, symbolic rules, graphs, or memory structures—that enables an AI system to interpret inputs, reason, make predictions, and perform tasks.

The Meta-Definition

A useful way to unify these concepts is as a layered model:

Reality → Observation → Data → Information → Semantics → Knowledge → Reasoning → Intelligence → Action

• Reality is what exists.

• Observation captures aspects of reality.

• Data records observations.

• Information organizes data to reduce uncertainty.

• Semantics gives information meaning.

• Knowledge connects meaningful information through validated relationships.

• Reasoning derives new conclusions from knowledge.

• Intelligence uses reasoning to adapt and achieve goals.

• Action applies intelligence to affect the world.

high-information, semantically explicit relationships 

Level 2 — Identity Relations

These help machines recognize that different references may denote the same thing.

• hasIdentifier

• hasURI

• hasUUID

• hasAlias

• hasSynonym

• hasCanonicalName

• sameAs

• equivalentTo

• disambiguatedBy

• references

Example:

OpenAI

    hasAlias

        "Open AI"

OpenAI

    hasWebsite

        https://openai.com

Level 3 — Classification Relations

These establish semantic hierarchy.

• subclassOf

• superclassOf

• categoryOf

• classifiedAs

• specializationOf

• generalizationOf

• siblingOf

Example

Dog

subclassOf

Mammal

Level 4 — Property Relations

Describe characteristics of entities.

• hasProperty

• hasAttribute

• hasValue

• hasUnit

• hasRange

• hasConstraint

Example

Earth

hasMass

5.97×10²⁴ kg

Level 5 — Temporal Relations

Represent time and change.

• createdOn

• updatedOn

• modifiedOn

• startedOn

• endedOn

• before

• after

• during

• simultaneousWith

• validUntil

Level 6 — Spatial Relations

Model physical or logical location.

• northOf

• southOf

• inside

• outside

• adjacentTo

• intersects

• overlaps

• surrounds

• nearestTo

Level 7 — Causal Relations

Essential for reasoning.

• causes

• causedBy

• enables

• prevents

• influences

• increases

• decreases

• triggers

• predicts

Example

Smoking

causes

Lung Cancer

Level 8 — Functional Relations

Capture behavior and purpose.

• performs

• performsOn

• executes

• computes

• generates

• transforms

• consumes

• produces

• controls

Level 9 — Scientific Relations

Useful across research domains.

• measuredBy

• observedBy

• discoveredBy

• provenBy

• validatedBy

• hypothesizedBy

• modeledBy

• simulatedBy

Level 10 — AI-Specific Relations

Critical for intelligent systems.

• representedByVector

• embeddedAs

• indexedIn

• retrievedBy

• rankedBy

• inferredFrom

• learnedFrom

• predictedBy

• classifiedBy

• reasonedFrom

Example

Entity

representedByVector

Embedding

Level 11 — Knowledge Graph Relations

Describe graph structure itself.

• connectedTo

• linkedTo

• neighborOf

• centralNode

• bridgeNode

• leafNode

• hubOf

• edgeWeight

• confidenceScore

Level 12 — Provenance Relations

Support trust and explainability.

• sourcedFrom

• verifiedBy

• citedBy

• supportedBy

• contradictedBy

• confidence

• evidence

Level 13 — Language Relations

Handle multilingual knowledge.

• translatedAs

• synonymOf

• antonymOf

• abbreviationOf

• acronymOf

• spellingVariantOf

Level 14 — Machine Learning Relations

Describe model behavior.

• trainedOn

• fineTunedOn

• evaluatedOn

• optimizedFor

• benchmarkedAgainst

• outperforms

• underperforms

Level 15 — Agent Relations

Model autonomous systems.

• knows

• remembers

• forgets

• plans

• decides

• requests

• delegates

• observes

• communicatesWith

• collaboratesWith

Level 16 — Human Knowledge Relations

Capture conceptual structure.

• defines

• explains

• illustrates

• compares

• contrasts

• extends

• summarizes

• derivesFrom

• exemplifies

Level 17 — Internet Relations

Represent the modern web.

• linksTo

• references

• cites

• publishes

• hosts

• mirrors

• indexes

• crawls

• serves

Level 18 — Business Relations

Support organizational knowledge.

• owns

• acquired

• founded

• funded

• partnersWith

• competesWith

• employs

• reportsTo

• manages

Level 19 — Research Relations

Model scientific knowledge.

• investigates

• proposes

• validates

• disproves

• reproduces

• compares

• extends

• surveys

Level 20 — Meta-Knowledge Relations

Describe knowledge about knowledge.

• describes

• represents

• abstracts

• formalizes

• annotates

• indexes

• organizes

• structures

• summarizes

• contextualizes

The AI-Optimized Semantic Hierarchy

The most valuable relationships are those that maximize semantic connectivity and support multiple reasoning tasks. A practical priority order is:

1. Identity — What is it?

2. Classification — What kind of thing is it?

3. Composition — What is it made of?

4. Properties — What characteristics does it have?

5. Location — Where does it exist?

6. Time — When does it exist or change?

7. Causality — Why does it happen?

8. Function — What does it do?

9. Interaction — How does it affect other entities?

10. Evidence — How do we know this?

11. Representation — How is it encoded for machines?

12. Reasoning — What can be inferred from it?