Knowledge Graph Schema¶

Overview¶

The knowledge graph uses a directed graph model with 5 node types and 9 edge types, designed to capture the relationships between Nobel laureates, their publications, technical concepts, and cross-disciplinary inspiration.

Node Types¶

Laureate¶

Represents a Nobel Prize laureate.

┌─────────────────┐
│    Laureate      │
│─────────────────│
│ id: string       │  "laureate_779"
│ type: "Laureate" │
│ name: string     │  "Aaron Ciechanover"
│ nationality: str │  "Israeli"
│ birth_year: int  │  1947
│ gender: string   │  "male"
└─────────────────┘

Award¶

Represents a specific Nobel Prize award event.

┌─────────────────┐
│      Award       │
│─────────────────│
│ id: string       │  "award_2004_3_779"
│ type: "Award"    │
│ year: int        │  2004
│ category: string │  "Chemistry"
│ motivation: str  │  "for the discovery of ubiquitin-mediated..."
│ prize_amount: int│
└─────────────────┘

Work¶

Represents a scientific publication.

┌──────────────────┐
│      Work         │
│──────────────────│
│ id: string        │  "W2078536640"
│ type: "Work"      │
│ title: string     │  "The ubiquitin-proteasome..."
│ year: int         │  1998
│ abstract: string  │
│ keywords: string  │
│ citation_count: int│  1250
│ doi: string       │
└──────────────────┘

Concept¶

Represents a technical or scientific concept extracted from publications.

┌────────────────────┐
│     Concept         │
│────────────────────│
│ id: string          │  "concept_ubiquitin_proteasome_pathway"
│ type: "Concept"     │
│ name: string        │  "Ubiquitin-Proteasome Pathway"
│ field: string       │  "Biology"
│ subfield: string    │  "Molecular Biology"
│ confidence: float   │  0.95
│ first_appeared: int │  1980
└────────────────────┘

Field¶

Represents a scientific discipline or domain.

┌────────────────────┐
│     Field           │
│────────────────────│
│ id: string          │  "field_biology"
│ type: "Field"       │
│ name: string        │  "Biology"
│ parent_field: str   │  null
└────────────────────┘

Edge Types¶

Relationship Map¶

graph LR
    L[Laureate] -->|WON_AWARD| A[Award]
    L -->|AUTHORED| W[Work]
    W -->|CITES| W2[Work]
    W -->|INTRODUCES| C[Concept]
    W -->|APPLIES| C
    C -->|BELONGS_TO| F[Field]
    C -->|DERIVED_FROM| C2[Concept]
    C -->|CROSS_INSPIRED| C3[Concept]
    C -->|ENABLED| C4[Concept]
    A -->|AWARDED_FOR| C

Edge Definitions¶

`WON_AWARD`¶

Property	Value
Source	Laureate
Target	Award
Attributes	`year`, `portion`
Semantics	Laureate received this Nobel Prize

`AUTHORED`¶

Property	Value
Source	Laureate
Target	Work
Attributes	`position`
Semantics	Laureate authored this paper

`CITES`¶

Property	Value
Source	Work
Target	Work
Attributes	—
Semantics	Paper A cites Paper B

`INTRODUCES`¶

Property	Value
Source	Work
Target	Concept
Attributes	`confidence`
Semantics	Paper first proposed or introduced this concept

`APPLIES`¶

Property	Value
Source	Work
Target	Concept
Attributes	`confidence`
Semantics	Paper applied or utilized this concept

`BELONGS_TO`¶

Property	Value
Source	Concept
Target	Field
Attributes	—
Semantics	Concept belongs to this scientific field

`DERIVED_FROM`¶

Property	Value
Source	Concept
Target	Concept
Attributes	`year`, `description`
Semantics	Concept evolved from another within the same field

`CROSS_INSPIRED` ⭐¶

Property	Value
Source	Concept
Target	Concept
Attributes	`year`, `source_field`, `target_field`, `description`
Semantics	Cross-disciplinary migration — a concept from one field inspired a concept in another

This is the core edge type of the knowledge graph. Examples:

Source	Target	Migration
Optimization Theory	Stochastic Gradient Descent	Math → AI (~1960s)
Transformer	AlphaFold	AI → Structural Biology (2018)
X-ray Diffraction	DNA Double Helix	Physics → Molecular Biology (1953)
Statistical Mechanics	Boltzmann Machine	Physics → Machine Learning (1985)
Quantum Mechanics	Quantum Chemistry	Physics → Chemistry (1930s)

`ENABLED`¶

Property	Value
Source	Concept
Target	Concept
Attributes	`description`
Semantics	One concept enabled or made possible another

`AWARDED_FOR`¶

Property	Value
Source	Award
Target	Concept
Attributes	—
Semantics	Nobel Prize was awarded for work on this concept

JSON Serialization¶

Full Graph Format¶

{
  "nodes": [
    {
      "id": "laureate_779",
      "type": "Laureate",
      "name": "Aaron Ciechanover",
      "nationality": "Israeli",
      "birth_year": 1947,
      "gender": "male"
    },
    {
      "id": "concept_ubiquitin",
      "type": "Concept",
      "name": "Ubiquitin-Proteasome Pathway",
      "field": "Biology",
      "subfield": "Molecular Biology"
    }
  ],
  "edges": [
    {
      "source": "laureate_779",
      "target": "award_2004_3_779",
      "type": "WON_AWARD",
      "year": 2004
    },
    {
      "source": "concept_a",
      "target": "concept_b",
      "type": "CROSS_INSPIRED",
      "year": 2001,
      "source_field": "Physics",
      "target_field": "Biology",
      "description": "Spectroscopy techniques applied to protein analysis"
    }
  ]
}

GraphML Export¶

The graph is also exported as GraphML (knowledge_graph.graphml) compatible with:

Gephi — Open-source graph visualization
Cytoscape — Network analysis platform
yEd — Graph editor
NetworkX — Python graph library

Concept Graph Schema¶

The Concept Graph is a simplified representation of the knowledge graph, focusing on concepts and their relationships. It is designed to highlight the flow of ideas and their connections across disciplines.

Schema Details¶

Nodes:
id: Unique identifier for the concept.
name: Human-readable name of the concept.
paper_count: Number of papers associated with the concept.
total_citations: Total citations received by papers linked to the concept.
Edges:
source: Source concept ID.
target: Target concept ID.
type: Relationship type (e.g., CONCEPT_CITES).
total_citations: Total citations between the connected concepts.

Construction Process¶

Extract concepts from papers.
Deduplicate concepts across papers.
Establish relationships based on citations and shared concepts.
Export the graph in JSON and GraphML formats.

Example JSON Structure¶

{
  "nodes": [
    {
      "id": "concept_1",
      "name": "Quantum Mechanics",
      "paper_count": 120,
      "total_citations": 4500
    }
  ],
  "edges": [
    {
      "source": "concept_1",
      "target": "concept_2",
      "type": "CONCEPT_CITES",
      "total_citations": 300
    }
  ]
}