Cite XWhy and SMILE¶
Cite the software artefact you used and the research method that supports your analysis. These are related but distinct citations.
Citation decision table¶
| What you used or discussed | Recommended research citation |
|---|---|
| General SMILE method, local statistical weighting, or classifier explanations | Foundational SMILE paper |
| Point-cloud explanations | Foundational SMILE paper + point-cloud SMILE preprint |
| Instruction-based image-editing explanations | Foundational SMILE paper + image-editing SMILE preprint |
| Large-language-model prompt explanations | Foundational SMILE paper + gSMILE preprint |
| Knowledge-graph or GraphRAG explanations | Foundational SMILE paper + KG-SMILE preprint |
| Concept-based explanation auditing | Foundational SMILE paper + ConceptSMILE preprint |
| Financial-sentiment reasoning with local perturbation explanations | gSMILE preprint + financial-sentiment study |
| The consolidated MSc study of generative-AI explainability | MSc thesis; add the relevant method paper where applicable |
Foundational SMILE citation¶
Use this as the primary methodological citation when referring to SMILE generally:
Koorosh Aslansefat, Mojgan Hashemian, Martin Walker, Mohammed Naveed Akram, Ioannis Sorokos, and Yiannis Papadopoulos. “Explaining Black Boxes with a SMILE: Statistical Model-Agnostic Interpretability with Local Explanations.” IEEE Software, 41(1), 87–97, 2024. https://doi.org/10.1109/MS.2023.3321282
@article{aslansefat2024smile,
author = {Koorosh Aslansefat and Mojgan Hashemian and Martin Walker
and Mohammed Naveed Akram and Ioannis Sorokos
and Yiannis Papadopoulos},
title = {Explaining Black Boxes with a {SMILE}: Statistical
Model-Agnostic Interpretability with Local Explanations},
journal = {IEEE Software},
year = {2024},
volume = {41},
number = {1},
pages = {87--97},
month = {Jan.--Feb.},
publisher = {IEEE},
doi = {10.1109/MS.2023.3321282},
url = {https://doi.org/10.1109/MS.2023.3321282}
}
Modality-specific citations¶
Point clouds¶
Cite the foundational SMILE paper and:
Seyed Mohammad Ahmadi, Koorosh Aslansefat, Rubén Valcarce-Diñeiro, and Joshua Barnfather. “Explainability of Point Cloud Neural Networks Using SMILE: Statistical Model-Agnostic Interpretability with Local Explanations.” arXiv:2410.15374, 2024. https://arxiv.org/abs/2410.15374
Instruction-based image editing¶
Cite the foundational SMILE paper and:
Zeinab Dehghani, Koorosh Aslansefat, Adil Khan, Adín Ramírez Rivera, Franky George, and Muhammad Khalid. “Mapping the Mind of an Instruction-Based Image Editing Using SMILE.” arXiv:2412.16277, 2024. https://arxiv.org/abs/2412.16277
Large language models¶
Cite the foundational SMILE paper and:
Zeinab Dehghani, Mohammed Naveed Akram, Koorosh Aslansefat, Adil Khan, and Yiannis Papadopoulos. “Explaining Large Language Models with gSMILE.” arXiv:2505.21657, 2025. https://arxiv.org/abs/2505.21657
Knowledge-graph retrieval-augmented generation¶
Cite the foundational SMILE paper and:
Zahra Zehtabi Sabeti Moghaddam, Zeinab Dehghani, Maneeha Rani, Koorosh Aslansefat, Bhupesh Kumar Mishra, Rameez Raja Kureshi, and Dhavalkumar Thakker. “Explainable Knowledge Graph Retrieval-Augmented Generation (KG-RAG) with KG-SMILE.” arXiv:2509.03626, 2025. https://arxiv.org/abs/2509.03626
Concept-based explainability¶
Cite the foundational SMILE paper and:
Mohadeseh Mollapour, Koorosh Aslansefat, Zeinab Dehghani, Bhupesh Kumar Mishra, Tejal Shah, and Zhibao Mian. “ConceptSMILE: Auditing the Trustworthiness of Concept-Based Explainable AI.” arXiv:2607.09649, 2026. https://arxiv.org/abs/2607.09649
Financial-sentiment reasoning¶
For the applied financial-sentiment study, cite gSMILE and:
Sania Verma, Koorosh Aslansefat, Joyjit Chatterjee, Akash Marar, Anu Mehra, and Aisha Ekundayo. “When Words Move Markets: Interpretable Behavioural and Robustness Analysis of LLMs for Financial Sentiment Reasoning via Local Perturbation Explanations.” In Natural Language Processing and Information Systems: NLDB 2026, LNCS 16696, 271–286. Springer, 2027; first published online 4 July 2026. https://doi.org/10.1007/978-3-032-29532-3_19
The publication catalogue provides complete BibTeX records for every method and study.
Citing the XWhy software¶
A research-method citation does not identify the exact software revision used in an experiment. Until the repository provides a versioned archived software release and formal software DOI, record:
- the package version;
- the Git commit SHA when reproducibility is important;
- the repository URL;
- the date on which the software was accessed;
- the configuration, model, data, and random seed used.
A temporary plain-text software reference can use this form:
XWhy contributors. XWhy: Model-Agnostic Explainability with SMILE. Version or commit used. GitHub repository: https://github.com/Dependable-Intelligent-Systems-Lab/xwhy. Accessed YYYY-MM-DD.
Citation principles¶
- Cite the foundational method when the work depends on SMILE’s statistical local-surrogate formulation.
- Add the relevant extension when the explained system is a point cloud, image editor, LLM, KG-RAG pipeline, or concept-based model.
- Distinguish preprints from peer-reviewed publications in prose and bibliographies.
- List the thesis separately from journal and conference papers.
- Record the software revision independently from the paper citation so an experiment can be reproduced.