CNCGEN: A Dataset and Framework for Machining Process Planning and Toolpath Generation from B-rep Models

arXiv 2026 53.1 method

TLDR

CNCGEN provides a dataset and learning framework for generating machining process plans and toolpaths from B-rep models, improving workpiece geometry on synthetic and real CNC records.

Reasoning

The paper presents a substantial dataset with verified synthetic and real CNC flows, and a framework that couples planning with geometric verification. Its main weakness is the narrow scope of three-axis machining and limited details on baseline comparisons in the abstract.

Read-first score

Read-first score 53.1, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 49.

Recency 8%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Methodology quality 25%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=baseline,dataset,experiment,metric,result

Reproducibility 25%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=dataset

Topical relevance 42%

30.6

Uses existing LLM keyword relevance scores normalized to 0-100. AI for CAD,computer-aided design,neural CAD,generative CAD,parametric CAD,B-Rep,boundary representation,constructive solid geometry,CSG,sketch extrusion,CAD generation,CAD reconstruction,text-to-CAD,image-to-CAD,point cloud to CAD,CAD program

Field roles

FrontierMethodology anchor

Rank sensitivity

Stability: volatile; rank range: 42.

Keyword Scores

B-Rep

boundary representation

AI for CAD

computer-aided design

neural CAD

generative CAD

CAD generation

CAD program

parametric CAD

constructive solid geometry

CSG

sketch extrusion

CAD reconstruction

text-to-CAD

image-to-CAD

point cloud to CAD

Deep Analysis

Innovations

Formulates machining process planning and toolpath generation around persistent manufacturing objects, where object identity determines the operation target and the evolving stock state conditions toolpath generation.
Introduces CNCGEN-Dataset, containing approximately 50k geometrically verified synthetic machining flows and 800 held-out real CNC records, with aligned B-rep geometry, object-referenced operations, parameterized toolpaths, intermediate stock states, and verification outcomes.
Proposes the CNCGEN framework, which generates operations and toolpaths for selected objects step by step and updates a compact machining state to guide subsequent predictions.
Uses a learned surrogate verifier during training to provide material-removal feedback that links local predictions to their geometric consequences.

Methodology

CNCGEN is a learning framework for three-axis machining that generates operations and toolpaths for selected objects step by step while updating a compact machining state. It is trained and evaluated using CNCGEN-Dataset, which contains approximately 50k synthetic verified machining flows and 800 held-out real CNC records aligned with B-rep geometry, object-referenced operations, parameterized toolpaths, intermediate stock states, and verification outcomes. A learned surrogate verifier provides material-removal feedback during training, and experiments compare CNCGEN against adapted CNC generation baselines using resulting workpiece geometry, residual material, and overcut.

Key Results

Experiments on synthetic and held-out real CNC records show that CNCGEN improves the resulting workpiece geometry and reduces residual material and overcut compared with adapted CNC generation baselines.

Resources

Paper