paper2graph

Causal dependency graphs of ML algorithms, extracted by Claude Opus 4.6 from research papers and reference implementations. Pick an algorithm, switch between training and inference phases, and toggle between three abstraction levels — macro (4–7 nodes), meso (7–10 nodes), and micro (12–25 nodes) — produced from the same source under prompt control.

Node ontology

Input
Parameter
Mechanism
State
Aggregation
Decision
Output

Choose an algorithm

Classical ML

K-means
Classical ML
PCA
Classical ML
Ridge Regression
Classical ML
Decision Trees
Classical ML

Ensemble

Random Forest
Ensemble
XGBoost
Ensemble
LightGBM
Ensemble

Transformer

Attention
Transformer
BERT
Transformer
Encoder–Decoder
Transformer
Image Transformer
Transformer

Generative

Diffusion Network
Generative

KG-oriented

TREK
KG-oriented
Wikontic
KG-oriented