Dependable-Intelligent-Systems-Lab · koo-ec · Sep 24, 2026 · Sep 24, 2026
diff --git a/docs/research/blogs.md b/docs/research/blogs.md
@@ -0,0 +1,26 @@
+---
+title: Research Blogs and Explainers
+description: Published SMILE and responsible-AI blog posts, plus XWhy guides about local explanations, LLMs, and evaluating attribution.
+---
+
+# Research blogs
+
+Explore published articles from the University of Hull research group and longer XWhy explainers. For research citations, use the [publication catalogue](publications.md).
+
+## Published research group blog posts
+
+- **[Mapping the Mind of an Instruction-Based Image Editing Using SMILE](https://www.responsibleaihull.com/post/mapping-the-mind-of-an-instruction-based-image-editing-using-smile)** — Koorosh Aslansefat explains how changes to words in an editing instruction can be connected to changes in the generated image.
+- **[Pint of Science: Can we trust AI for skin cancer? Are they fair?](https://www.responsibleaihull.com/post/pint-of-science-trustworthy-ai-can-we-trust-ai-for-skin-cancer-are-they-fair)** — A public discussion of fairness and reliability in medical image classification, featuring Koorosh Aslansefat and Kuniko Paxton.
+- **[Research Visits of Responsible AI Team to Chennai, India](https://www.responsibleaihull.com/post/research-visits-of-responsible-ai-team-to-chennai-india)** — A report on research exchanges about trustworthy AI in healthcare.
+- **[Responsible Digital Twin Research Theme in Chennai, India](https://www.responsibleaihull.com/post/responsible-digital-twin-research-theme-in-chennai-india)** — A related research visit on digital twins and dependable intelligent systems.
+
+[Browse all posts from the Centre for Responsible AI at Hull](https://www.responsibleaihull.com/research-blog).
+
+## XWhy explainers and practical reading
+
+- **[How SMILE works](../concepts/smile.md)** — The perturbation, weighting, and local surrogate workflow behind an explanation.
+- **[What a local explanation tells you](../concepts/local-explanations.md)** — How to interpret results for a particular input and model output.
+- **[Explaining LLM responses](../explainers/llm/index.md)** — Prompt perturbations and token contributions in language-model workflows.
+- **[Image generation and editing examples](../explainers/image-generation/image-generation-editing-examples.md)** — Examples and scope for visual generation workflows.
+- **[Evaluating explanation fidelity](../evaluation/attribution-fidelity.md)** — Ways to test whether an attribution reflects the model's behaviour.
+- **[Limitations and responsible interpretation](../concepts/limitations.md)** — Where local explanations need additional checks.
diff --git a/docs/research/index.md b/docs/research/index.md
@@ -17,6 +17,7 @@ SMILE—**Statistical Model-Agnostic Interpretability with Local Explanations**
 | Language and generative AI | gSMILE | Large language models | arXiv preprint |
 | Retrieval-augmented generation | KG-SMILE | Knowledge-graph and GraphRAG systems | arXiv preprint |
 | Concept-based explainability | ConceptSMILE | Concept-based explainable-AI methods | arXiv preprint |
+| Concept-based explainability | ConceptSMILE thesis | Retinal-image AI and trustworthy visual concepts | MSc thesis |
 | Financial language analysis | Local perturbation explanations derived from gSMILE | LLM financial-sentiment reasoning | Peer-reviewed conference chapter |
 | Consolidated academic study | Generative-AI SMILE thesis | LLMs and instruction-based image editing | MSc thesis |
 
@@ -45,6 +46,7 @@ ConceptSMILE audits whether human-understandable concepts provide explanations t
 ## Research resources
 
 - [Browse the complete publication catalogue](publications.md)
+- [Read research blogs and explainers](blogs.md)
 - [Choose the correct citation for your use case](citation.md)
 - [Follow the reproducibility guidance](../how-to/reproducibility.md)
 

diff --git a/docs/research/publications.md b/docs/research/publications.md
@@ -17,6 +17,7 @@ This catalogue connects the SMILE research family to the systems each publicatio
 | 2025 | gSMILE | Large language models | arXiv preprint | [arXiv:2505.21657](https://arxiv.org/abs/2505.21657) |
 | 2025 | KG-SMILE | Knowledge-graph and GraphRAG systems | arXiv preprint | [arXiv:2509.03626](https://arxiv.org/abs/2509.03626) |
 | 2026 | ConceptSMILE | Concept-based explainable AI | arXiv preprint | [arXiv:2607.09649](https://arxiv.org/abs/2607.09649) |
+| 2026 | ConceptSMILE thesis | Retinal-image explainability and trustworthy visual concepts | MSc thesis | [University of Hull repository](https://hull-repository.worktribe.com/output/5763337) |
 | 2026/2027 | Financial-sentiment local perturbation study | LLM financial-sentiment reasoning | NLDB 2026 conference chapter; 2027 volume | [DOI](https://doi.org/10.1007/978-3-032-29532-3_19) |
 | 2025 | Generative-AI SMILE thesis | LLM and image-editing explainability | MSc thesis | [arXiv:2602.01206](https://arxiv.org/abs/2602.01206) |
 
@@ -226,6 +227,28 @@ ConceptSMILE audits whether higher-level, human-understandable concepts provide
 
 ## Related thesis
 
+### ConceptSMILE: Towards A Framework for Trustworthy Concept-Based Explainable Artificial Intelligence using SMILE
+
+**Mohadeseh Mollapour.** “ConceptSMILE: Towards A Framework for Trustworthy Concept-Based Explainable Artificial Intelligence using SMILE.” MSc thesis, University of Hull, 2026. Supervised by Koorosh Aslansefat and Bhupesh Mishra.
+
+This thesis studies concept-based explanations for retinal-image AI and evaluates their trustworthiness. It is a separate academic record from the [ConceptSMILE arXiv preprint](https://arxiv.org/abs/2607.09649) listed above.
+
+[Thesis record](https://hull-repository.worktribe.com/output/5763337)
+
+??? note "BibTeX"
+
+    ```bibtex
+    @mastersthesis{mollapour2026conceptsmilethesis,
+      author = {Mohadeseh Mollapour},
+      title  = {{ConceptSMILE}: Towards A Framework for Trustworthy
+                Concept-Based Explainable Artificial Intelligence using {SMILE}},
+      school = {University of Hull},
+      type   = {{MSc} thesis},
+      year   = {2026},
+      url    = {https://hull-repository.worktribe.com/output/5763337}
+    }
+    ```
+
 ### Addressing Explainability of Generative AI Using SMILE
 
 **Zeinab Dehghani.** “Addressing Explainability of Generative AI Using SMILE: Statistical Model-Agnostic Interpretability with Local Explanations.” MSc thesis, University of Hull, September 2025. Also available as arXiv:2602.01206.

diff --git a/properdocs.yml b/properdocs.yml
@@ -92,6 +92,7 @@ plugins:
           - evaluation/attribution-faithfulness.md: ATT faithfulness under interventions
         Research:
           - research/index.md: Research and reproducibility entry point
+          - research/blogs.md: Research blogs and XWhy explainers
           - research/citation.md: How to cite XWhy
         Releases:
           - release-notes.md: Version history and detailed release notes
@@ -159,6 +160,7 @@ nav:
   - Research:
       - Overview: research/index.md
       - Publications: research/publications.md
+      - Blogs: research/blogs.md
       - Citation: research/citation.md
   - API Reference: reference/index.md
   - Release Notes: release-notes.md