GraphSAGE
Architectures AIMech:GraphSAGE DRAFT
An inductive graph representation architecture that learns neighborhood aggregation functions to generate node embeddings from features.
Identity Notes & Concept Scope
GraphSAGE learns an embedding-generating function, unlike individually fitted transductive node embeddings. The broad AIO graph-convolutional category does not require the named sampling-and-aggregation framework.
Claims & Grounded Evidence
1 Claim
Mechanism
#mechanism
At each depth, sampled neighbor representations are aggregated and combined with a node's current representation through learned transformations; parameters are shared across nodes.
⚠️ Curated Limitations & Negative Results
- The paper describes multiple aggregators; no single mean or pooling variant defines the entire framework.
- Inductive construction does not guarantee accurate embeddings for every unseen graph or feature distribution.
Artificial Intelligence Ontology (AIO) Grounding
Mapped AIO Classes
| AIO Target Class | Mapping Relation | Mapping Rationale |
|---|---|---|
AIO:ArtificialNeuralNetwork
Artificial Neural Network
|
broad_match | Learned aggregation and transformation layers form an artificial neural network; this target is broader than the specific graph architecture. |
AIO:GraphConvolutionalNetwork
Graph Convolutional Network
|
related_match | The pinned definition broadly describes learning on graph structure, but GraphSAGE has its own inductive sampling/aggregation construction; a narrower GCN algorithm is not equated with it. |
Evidence Sources & Primary Literature
1 Source| Source ID | Publication / Document | Type / Access | Reference & Checksum |
|---|---|---|---|
original-paper
|
William L. Hamilton, Rex Ying, Jure Leskovec
|
primary paper
full_text
|
ID: arXiv:1706.02216v4
sha256:e1f354eb1201…
|
Curation Provenance
Curator: Codex
Method: ai_assisted
Model: GPT-6
Created: 2026-10-11T03:12:53Z
Updated: 2026-10-11T03:12:53Z
Source File:
data/curated/GraphSAGE.yaml