Cornus officinalis Sieb. et Zucc. is a medicinal and ornamental woody plant whose distribution and production may be increasingly affected by climate change. This study assessed its climatic suitability under paleoclimatic, current, and future conditions to provide a climatic basis for cultivation planning, introduction, and germplasm conservation. An optimized MaxEnt model was developed using 381 spatially filtered occurrence records and five environmental predictors. Model complexity was tuned using spatial block cross-validation, and future projections for 2081–2100 were generated from an ensemble of ACCESS-CM2, BCC-CSM2-MR, and CMCC-ESM2 under four Shared Socioeconomic Pathway scenarios. Cold-quarter temperature, warm-season precipitation, mean diurnal temperature range, precipitation seasonality, and slope jointly shaped the predicted distribution. Under the current climate, climatically suitable areas were concentrated mainly in East Asia, particularly central and eastern China, with additional suitable climatic regions in eastern North America and parts of Europe. Suitable area was smallest during the Last Interglacial, expanded during the Last Glacial Maximum, and approached its current extent during the Mid-Holocene. The current suitable area was estimated at 277.247 × 104 km2 and was projected to decrease to 243.689, 206.615, 201.393, and 206.168 × 104 km2 under SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5, respectively. Existing suitable areas contracted in central China, the southern United States, and southern Japan, whereas climatic suitability expanded northward in parts of China, North America, Japan, and Europe. Suitable-area gains did not offset losses under any scenario, and retention of current suitable areas declined from 52.88% under SSP1-2.6 to 19.74% under SSP5-8.5. Most newly suitable areas were supported by at least two of the three climate models, although greater inter-model variation occurred near some range margins. These findings indicate substantial future redistribution of climatic suitability and identify climatically stable and newly suitable regions that may inform climate-resilient cultivation planning and conservation of C. officinalis.
Ranking models are repeatedly refined within established model families, yet the choice of which component to scale, replace, or retire is often guided by expensive trial-and-error. We present Model Internal State Optimization (MISO), a systems workflow that uses model internal states (MIS), including parameters, activations, gradients, and normalization statistics, to prioritize such local optimization decisions. MISO extracts MIS from a trained ranking model, aggregates them into ranking, alignment, and comparison signals, and converts those signals into a small set of interpretable candidate edits. Because MIS are re-extracted after each retraining cycle, MISO naturally supports an adaptive optimization workflow that tracks evolving model behavior as data distributions and system requirements shift over time. In an ads ranking case study, MISO improves normalized entropy while requiring substantially fewer validation runs than expert-driven and black-box scaling workflows, offering a practical middle ground between manual tuning and opaque automated search.
Yongzhen Zhang, Xiaoyu Deng, Yifan He et al.· 0 citations