The development of smart coatings with potential for active anticorrosion and self-healing protection of metals is essential for long-term performance of metallic structures in aggressive chemical environments.Present...The development of smart coatings with potential for active anticorrosion and self-healing protection of metals is essential for long-term performance of metallic structures in aggressive chemical environments.Presently,emphasis has been placed on the development of advanced smart coatings for corrosion protection in different applications.Innovative multifunctional coatings with fascinating stimuliresponsive functionalities are considered“smart”.The stimuli-responsive functionalities of these smart coatings when properly harnessed result in a class of coatings with inherent autonomous control of corrosion.Fundamentally,when metals are exposed to aggressive environments,occurrences at the metalsolution interface cause environmental changes.These changes can be controlled when triggers from external environment set off active components of smart coating,thereby enhancing coating’s life and functionality.Common triggers include the availability of moisture,concentration of chloride ion,p H gradient,mechanical damage,impact,fatigue,light,redox activity and temperature.In this review,recent technological trends in active anticorrosion and self-healing coatings as functional routes for metal protection are summarized,stimuli responsiveness and mechanisms of inhibition are discussed,and recent multi-action protective systems are particularly focused on.展开更多
预训练世界模型是提升强化学习样本效率的关键技术,但现有方法因视频数据缺乏显式动作标注,难以捕捉状态转移的因果机制。对此,提出多模态大模型辅助的视频动作生成预训练框架(MLM-generated Action-based Pre-training from videos for...预训练世界模型是提升强化学习样本效率的关键技术,但现有方法因视频数据缺乏显式动作标注,难以捕捉状态转移的因果机制。对此,提出多模态大模型辅助的视频动作生成预训练框架(MLM-generated Action-based Pre-training from videos for world models,MAPO),通过整合视觉语言模型的语义理解能力与动力学建模需求,突破传统预训练范式在动作语义缺失方面的局限性。具体地,MAPO在预训练阶段利用多模态大模型(QWEN2_5-VL-7B)解析视频帧序列,生成细粒度语义动作描述,构建具有因果解释性的动作-状态关联;设计上下文量化编码机制,解耦场景静态特征与动态控制因素,增强跨模态表征能力。在微调阶段,通过双网络协同架构实现预训练动力学特征与真实环境动作的端到端对齐。实验表明,MAPO在DeepMind Control Suite和Meta-World的8项任务中的平均回报较最优基线获得稳定提升,尤其在长时程任务中展现出卓越的性能。该研究为跨模态世界模型训练提供了新范式,揭示了语义动作生成在因果推理中的关键作用。展开更多
The lithium-sulfur battery has attracted enormous attention as being one of the most significant energy storage technologies due to its high energy density and cost-effectiveness.However,the "shuttle effect"...The lithium-sulfur battery has attracted enormous attention as being one of the most significant energy storage technologies due to its high energy density and cost-effectiveness.However,the "shuttle effect" of polysulfide intermediates represents a formidable challenge towards its wide applications.Herein,we have designed and synthesized two-dimensional Cu,Zn and Sn-based multimetallic sulfide nanosheets to construct multi-active sites for the immobilization and entrapment of polysulfides with offering better performance in liquid Li2S6-based lithium-polysulfide batteries.Both experimental measurements and theoretical computations demonstrate that the interfacial multi-active sites of multimetallic sulfides not only accelerate the multi-chained redox reactions of highly diffusible polysulfides,but also strengthen affinities toward polysulfides.By adopting multimetallic sulfide nanosheets as the sulfur host,the liquid Li2 S6-based cell exhibits an impressive rate capability with 1200 mAh/g and retains 580 mAh/g at 0.5 mA/cm^(2) after 1000 cycles.With high sulfur mass loading conditions,the cell with 2.0 mg/cm^(2) sulfur loading delivers a cell capacity of 1068 mAh/g and maintains 480 mAh/g with 0.8 mA/cm^(2) and 500 cycles.This study provides new insights into the multifunctional material design with multi-active sites for elevated lithium-polysulfide batteries.展开更多
基金financial support from the National Natural Science Foundation of China(Nos.52171089 and 51571202)Ling Chuang Research Project of China National Nuclear Corporation。
文摘The development of smart coatings with potential for active anticorrosion and self-healing protection of metals is essential for long-term performance of metallic structures in aggressive chemical environments.Presently,emphasis has been placed on the development of advanced smart coatings for corrosion protection in different applications.Innovative multifunctional coatings with fascinating stimuliresponsive functionalities are considered“smart”.The stimuli-responsive functionalities of these smart coatings when properly harnessed result in a class of coatings with inherent autonomous control of corrosion.Fundamentally,when metals are exposed to aggressive environments,occurrences at the metalsolution interface cause environmental changes.These changes can be controlled when triggers from external environment set off active components of smart coating,thereby enhancing coating’s life and functionality.Common triggers include the availability of moisture,concentration of chloride ion,p H gradient,mechanical damage,impact,fatigue,light,redox activity and temperature.In this review,recent technological trends in active anticorrosion and self-healing coatings as functional routes for metal protection are summarized,stimuli responsiveness and mechanisms of inhibition are discussed,and recent multi-action protective systems are particularly focused on.
文摘预训练世界模型是提升强化学习样本效率的关键技术,但现有方法因视频数据缺乏显式动作标注,难以捕捉状态转移的因果机制。对此,提出多模态大模型辅助的视频动作生成预训练框架(MLM-generated Action-based Pre-training from videos for world models,MAPO),通过整合视觉语言模型的语义理解能力与动力学建模需求,突破传统预训练范式在动作语义缺失方面的局限性。具体地,MAPO在预训练阶段利用多模态大模型(QWEN2_5-VL-7B)解析视频帧序列,生成细粒度语义动作描述,构建具有因果解释性的动作-状态关联;设计上下文量化编码机制,解耦场景静态特征与动态控制因素,增强跨模态表征能力。在微调阶段,通过双网络协同架构实现预训练动力学特征与真实环境动作的端到端对齐。实验表明,MAPO在DeepMind Control Suite和Meta-World的8项任务中的平均回报较最优基线获得稳定提升,尤其在长时程任务中展现出卓越的性能。该研究为跨模态世界模型训练提供了新范式,揭示了语义动作生成在因果推理中的关键作用。
基金supported by the Start-up Foundation of Nanjing Tech Universitythe National Natural Science Foundation of China (61904080, 61801210, 91833302)+3 种基金the Natural Science Foundation of Jiangsu Province (BK20190670, BK20180686)the Natural Science Foundation of Colleges and Universities in Jiangsu Province (19KJB530008)the Innovation Scientists and Technicians Team Construction Projects of Henan Province (CXTD2017002)the funding for “Distinguished professors” and “High-level talents in six industries” of Jiangsu Province and Technology Innovation Project for Overseas Scholar in Nanjing。
文摘The lithium-sulfur battery has attracted enormous attention as being one of the most significant energy storage technologies due to its high energy density and cost-effectiveness.However,the "shuttle effect" of polysulfide intermediates represents a formidable challenge towards its wide applications.Herein,we have designed and synthesized two-dimensional Cu,Zn and Sn-based multimetallic sulfide nanosheets to construct multi-active sites for the immobilization and entrapment of polysulfides with offering better performance in liquid Li2S6-based lithium-polysulfide batteries.Both experimental measurements and theoretical computations demonstrate that the interfacial multi-active sites of multimetallic sulfides not only accelerate the multi-chained redox reactions of highly diffusible polysulfides,but also strengthen affinities toward polysulfides.By adopting multimetallic sulfide nanosheets as the sulfur host,the liquid Li2 S6-based cell exhibits an impressive rate capability with 1200 mAh/g and retains 580 mAh/g at 0.5 mA/cm^(2) after 1000 cycles.With high sulfur mass loading conditions,the cell with 2.0 mg/cm^(2) sulfur loading delivers a cell capacity of 1068 mAh/g and maintains 480 mAh/g with 0.8 mA/cm^(2) and 500 cycles.This study provides new insights into the multifunctional material design with multi-active sites for elevated lithium-polysulfide batteries.
