← Back to the wire

iLENS: Interpretable LLM-Guided Mixture-of-Experts for Neuroimaging Survival Analysis

AnnouncementResearchJun 12, 2026

iLENS is an LLM-guided mixture-of-experts framework for survival prediction in Alzheimer's disease conversion. Proposed by Farica Zhuang, the framework synthesizes structured neuroimaging measurements and unstructured information to guide expert routing. iLENS demonstrates competitive predictive performance and patient subtyping capability while providing "transparent, biologically grounded rationales" for its routing decisions, addressing the limited interpretability of typical static survival models used in AD risk prediction.

Receipt № 5951 source · awaiting confirmation ◐

Evidence

1source· awaiting independent confirmation

No score is assigned. Sources and their independence are shown in the citation chain below.

Citation chain · 1 source

01medPRIMARY
iLENSModelFarica ZhuangPerson
Canonical: https://arxiv.org/abs/2607.08778