高效准确的机器学习模型是高通量筛选优质含能分子的基础,为此设计了完全通过结构式便可输出多项含能材料性质的多输入多输出机器学习模型(MIMO-ML)。基于478个含能分子数据集进行模型构建,通过筛选得到AMID_h、ATTS1are、OB三个优秀描...高效准确的机器学习模型是高通量筛选优质含能分子的基础,为此设计了完全通过结构式便可输出多项含能材料性质的多输入多输出机器学习模型(MIMO-ML)。基于478个含能分子数据集进行模型构建,通过筛选得到AMID_h、ATTS1are、OB三个优秀描述符,并在此基础上建立了自定义描述符集。结果表明,随机森林(RF)及多层感知机(MLP)适合被应用于含能材料性能预测多输出模型构建,输出性质包括爆速D(MAE=256 m/s),热分解温度T_(d)(MAE=34.7℃),撞击感度ln H 50(MAE=0.63)。同时,MLP模型相比于RF模型对于特征数量更敏感,且利用更少的特征可得到与RF精度相似的模型,表明MIMO-ML模型能够快速且准确地识别高性能含能材料,可应用于含能分子的设计与快速筛选。展开更多
This study explores the motivations,perceived benefits,and challenges associated with the adoption of clearcutfree forestry by early adopters among non-industrial private forest(NIPF)owners in southern-central Sweden....This study explores the motivations,perceived benefits,and challenges associated with the adoption of clearcutfree forestry by early adopters among non-industrial private forest(NIPF)owners in southern-central Sweden.Clearcut-free forestry,characterized by continuous tree cover and an emphasis on biodiversity,structural diversity,and ecosystem services(ES),is increasingly seen as a sustainable alternative to conventional intensive management based on short rotations and clear-cutting practices.Based on qualitative interviews with 22 NIPF owners who have adopted this approach,the study provides insights into how these early adopters perceive the value of clearcut-free forestry.Reported motivations include environmental concerns,such as biodiversity conservation and climate resilience,as well as strong socio-cultural values linked to family traditions,aesthetic preferences,and community wellbeing.In this study,we use the multi-level perspective(MLP)framework to conceptualize NIPF owners who have adopted clearcut-free forestry as niche actors and analyze their potential contribution to the emergence of an alternative forest management regime.The findings highlight that early adopters associate multiple benefits with clearcut-free forestry,encompassing enhanced ecosystem services such as carbon sequestration,water regulation,habitat preservation,and socio-cultural enrichment through recreation and relational values.However,the interviewees identify several interrelated challenges,including knowledge gaps,lack of clear definitions and standardized practices,limited advisory services,underdeveloped value chains for high-quality timber,and market barriers,which hinder more widespread adoption.Within the multi-level perspective,owner perceptions linking clearcut-free management with improved forest multifunctionality serve as a key driver of niche-level experimentation.This suggests an alignment between these owners and evolving societal demands for more inclusive,sustainable,and diversified forest use.Policy recommendations include targeted investments in knowledge co-production,infrastructure,market incentives,and certification schemes to support the economic viability and broader adoption of clearcut-free forestry.展开更多
文摘高效准确的机器学习模型是高通量筛选优质含能分子的基础,为此设计了完全通过结构式便可输出多项含能材料性质的多输入多输出机器学习模型(MIMO-ML)。基于478个含能分子数据集进行模型构建,通过筛选得到AMID_h、ATTS1are、OB三个优秀描述符,并在此基础上建立了自定义描述符集。结果表明,随机森林(RF)及多层感知机(MLP)适合被应用于含能材料性能预测多输出模型构建,输出性质包括爆速D(MAE=256 m/s),热分解温度T_(d)(MAE=34.7℃),撞击感度ln H 50(MAE=0.63)。同时,MLP模型相比于RF模型对于特征数量更敏感,且利用更少的特征可得到与RF精度相似的模型,表明MIMO-ML模型能够快速且准确地识别高性能含能材料,可应用于含能分子的设计与快速筛选。
基金financed by a grant from Mistra[DIA 2019/28]and from Formas via the National Research Programme on Climate(2021–00416)FORMAS,Grant Nos.2022-02146 and 2021–01067Swedish Environmental Protection Agency Research Grant No.2021–00040。
文摘This study explores the motivations,perceived benefits,and challenges associated with the adoption of clearcutfree forestry by early adopters among non-industrial private forest(NIPF)owners in southern-central Sweden.Clearcut-free forestry,characterized by continuous tree cover and an emphasis on biodiversity,structural diversity,and ecosystem services(ES),is increasingly seen as a sustainable alternative to conventional intensive management based on short rotations and clear-cutting practices.Based on qualitative interviews with 22 NIPF owners who have adopted this approach,the study provides insights into how these early adopters perceive the value of clearcut-free forestry.Reported motivations include environmental concerns,such as biodiversity conservation and climate resilience,as well as strong socio-cultural values linked to family traditions,aesthetic preferences,and community wellbeing.In this study,we use the multi-level perspective(MLP)framework to conceptualize NIPF owners who have adopted clearcut-free forestry as niche actors and analyze their potential contribution to the emergence of an alternative forest management regime.The findings highlight that early adopters associate multiple benefits with clearcut-free forestry,encompassing enhanced ecosystem services such as carbon sequestration,water regulation,habitat preservation,and socio-cultural enrichment through recreation and relational values.However,the interviewees identify several interrelated challenges,including knowledge gaps,lack of clear definitions and standardized practices,limited advisory services,underdeveloped value chains for high-quality timber,and market barriers,which hinder more widespread adoption.Within the multi-level perspective,owner perceptions linking clearcut-free management with improved forest multifunctionality serve as a key driver of niche-level experimentation.This suggests an alignment between these owners and evolving societal demands for more inclusive,sustainable,and diversified forest use.Policy recommendations include targeted investments in knowledge co-production,infrastructure,market incentives,and certification schemes to support the economic viability and broader adoption of clearcut-free forestry.
