Dao-Danzi Germplasm Prototypes in Cave-Thermostat-Like Niches: A Genmount Ecological Predictive State-Space Model with Debiasing and Invariant Learning for Pharmacopoeial Quality Proxies and Semantic Explanations of TCM Properties

Abstract

Dao-di herbal quality arises from coupled microclimate buffering, community interactions, metabolic networks, and pharmacopoeial metrics. We propose DD-CHEI: a falsifiable multi-species platform that defines Dao-Danzi (DD) as enumerable germplasm prototypes and action sequences in cave-thermostat-like niches versus open controls; freezes CHEI microclimate indices, a 2x2 experimental design, mixed-effects / PERMANOVA plans, and a Genmount ecological predictive state-space model with adversarial debiasing, IRM, and philological weak supervision for a Siqi-Wuwei semantic explanation layer. This Version publishes methods and a frozen evaluation protocol; no measured CHEI effect sizes, quality-proxy contrasts, or LOPO/LOSO scores are claimed. A small helper (dd_chei_ref.py) is included. Chinese and English full texts are deposited with matching section numbering and markdown structure.