Tabular adversarial-ML research: strategy and evaluation taxonomy

Research applicability

A taxonomy and evaluation reference for tabular adversarial-ML research.

Getting started

Read the linked primary sources and the stated adoption requirements. Reuse suggestions are editorial and untested; no setup recipe or trial was executed.

Maintenance

Sources last checked on 2026-10-19. Future availability, maintenance, provider terms and compatibility were not verified.

License

Paper, figure and dataset rights were not audited.

Workflow & human oversight

No workflow level assigned. The review categorises tabular adversarial-ML method families and practical evaluation dimensions. Its human-authored categorisation does not establish one AI-assisted research workflow or validation protocol.

Workflow level is not an evidence score; higher is not better evidence.

Approach

Selected text maps strategy families and marks practical requirements as considered, acknowledged or absent. Discussion notes inconsistent definitions, limited benchmarks and uncertain dataset generalisation. The taxonomy is a reference, not a validated defence or attack implementation.

Limitations

Selected HTML method, taxonomy, evaluation and discussion passages only. The complete paper, appendices, primary studies, code, datasets and numerical findings were not audited. Firstness, completeness and attack-count claims are not validated. Feasibility, semantics and plausibility are inconsistently defined, so source categories do not prove real-world validity. No attack procedure, security-system change or experiment was performed.

Reproducibility

Source text was checked only within this scope: Selected primary HTML method/taxonomy/evaluation/discussion passages. No software, dataset, experiment or result was independently reproduced.

Adoption requirements

Untested suggestion: use the taxonomy to locate primary studies and check definitions, constraints, benchmarks and dataset transfer before research reuse.