ProfileXAI: User-Adaptive Explainable AI
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917044958527488 |
|---|---|
| author | Corrales, Gilber A. Sánchez, Carlos Andrés Ferro Tabares-Soto, Reinel Sotelo, Jesús Alfonso López Ruz, Gonzalo A. Durán, Johan Sebastian Piña |
| author_facet | Corrales, Gilber A. Sánchez, Carlos Andrés Ferro Tabares-Soto, Reinel Sotelo, Jesús Alfonso López Ruz, Gonzalo A. Durán, Johan Sebastian Piña |
| contents | ProfileXAI is a model- and domain-agnostic framework that couples post-hoc explainers (SHAP, LIME, Anchor) with retrieval - augmented LLMs to produce explanations for different types of users. The system indexes a multimodal knowledge base, selects an explainer per instance via quantitative criteria, and generates grounded narratives with chat-enabled prompting. On Heart Disease and Thyroid Cancer datasets, we evaluate fidelity, robustness, parsimony, token use, and perceived quality. No explainer dominates: LIME achieves the best fidelity-robustness trade-off (Infidelity $\le 0.30$, $L<0.7$ on Heart Disease); Anchor yields the sparsest, low-token rules; SHAP attains the highest satisfaction ($\bar{x}=4.1$). Profile conditioning stabilizes tokens ($σ\le 13\%$) and maintains positive ratings across profiles ($\bar{x}\ge 3.7$, with domain experts at $3.77$), enabling efficient and trustworthy explanations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_22998 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | ProfileXAI: User-Adaptive Explainable AI Corrales, Gilber A. Sánchez, Carlos Andrés Ferro Tabares-Soto, Reinel Sotelo, Jesús Alfonso López Ruz, Gonzalo A. Durán, Johan Sebastian Piña Artificial Intelligence 68T05, 68T07 I.2.6; H.5.2 ProfileXAI is a model- and domain-agnostic framework that couples post-hoc explainers (SHAP, LIME, Anchor) with retrieval - augmented LLMs to produce explanations for different types of users. The system indexes a multimodal knowledge base, selects an explainer per instance via quantitative criteria, and generates grounded narratives with chat-enabled prompting. On Heart Disease and Thyroid Cancer datasets, we evaluate fidelity, robustness, parsimony, token use, and perceived quality. No explainer dominates: LIME achieves the best fidelity-robustness trade-off (Infidelity $\le 0.30$, $L<0.7$ on Heart Disease); Anchor yields the sparsest, low-token rules; SHAP attains the highest satisfaction ($\bar{x}=4.1$). Profile conditioning stabilizes tokens ($σ\le 13\%$) and maintains positive ratings across profiles ($\bar{x}\ge 3.7$, with domain experts at $3.77$), enabling efficient and trustworthy explanations. |
| title | ProfileXAI: User-Adaptive Explainable AI |
| topic | Artificial Intelligence 68T05, 68T07 I.2.6; H.5.2 |
| url | https://arxiv.org/abs/2510.22998 |