针对于工业场所中密集场景下安全帽佩戴出现的漏检情况以及提高检测精度,提出了一种基于改进YOLOv5s的工业安全帽检测算法.首先,采用CIOU优化Soft-NMS对密集人群重叠的情况,减少重叠目标的漏检,从而提高了安全帽检测的准确性.其次,在网...针对于工业场所中密集场景下安全帽佩戴出现的漏检情况以及提高检测精度,提出了一种基于改进YOLOv5s的工业安全帽检测算法.首先,采用CIOU优化Soft-NMS对密集人群重叠的情况,减少重叠目标的漏检,从而提高了安全帽检测的准确性.其次,在网络的中间层添加辅助训练头引入丰富的梯度信息,最后,辅助训练头结合Optimal Transport Assignment添加到Loss中,通过最优的目标匹配,减少模型的漏检和误检的情况,从而提升模型的准确率和召回率.实验结果表明,改进后的算法平均精确值(mAP@50-90)值为68.3%,相对于原YOLOv5s算法提升了3.8%,准确率为92.3%,相较于原YOLOv5s算法提高了0.7%.展开更多
The development of chemical technologies,which involves a multistage process covering laboratory research,scale‐up to industrial deployment,and necessitates interdisciplinary collaboration,is often accompanied by sub...The development of chemical technologies,which involves a multistage process covering laboratory research,scale‐up to industrial deployment,and necessitates interdisciplinary collaboration,is often accompanied by substantial time and economic costs.To address these challenges,in this work,we report ChemELLM,a domain‐specific large language model(LLM)with 70 billion parameters for chemical engineering.ChemELLM demonstrates state‐of‐the‐art performance across critical tasks ranging from foundational understanding to professional problem‐solving.It outperforms mainstream LLMs(e.g.,O1‐Preview,GPT‐4o,and DeepSeek‐R1)on ChemEBench,the first multidimensional benchmark for chemical engineering,which encompasses 15 dimensions across 101 distinct essential tasks.To support robust model development,we curated ChemEData,a purpose‐built dataset containing 19 billion tokens for pre‐training and 1 billion tokens for fine‐tuning.This work establishes a new paradigm for artificial intelligence‐driven innovation,bridging the gap between laboratory‐scale innovation and industrial‐scale implementation,thus accelerating technological advancement in chemical engineering.ChemELLM is publicly available at https://chemindustry.iflytek.com/chat.展开更多
针对英语口语考试中常见的封闭类和开放类问答题,在中考口语模拟考试数据集合上,对比了基于特征工程、双向编码器小尺寸文本预训练模型(BERT)和单向解码器大尺寸文本预训练模型(Large Language Model,LLM)三种自动评分方法。实验结果表...针对英语口语考试中常见的封闭类和开放类问答题,在中考口语模拟考试数据集合上,对比了基于特征工程、双向编码器小尺寸文本预训练模型(BERT)和单向解码器大尺寸文本预训练模型(Large Language Model,LLM)三种自动评分方法。实验结果表明,不管是在封闭类还是开放类问答题型上,基于LLM的自动评分方法表现最优,BERT模型次之,特征工程方法最差。在开放类问答上,基于LLM和BERT的自动评分相对特征工程方案优势更为明显,证明LLM和BERT模型在问答题这种需要语义理解的任务上具备天然优势,为大模型技术应用到大规模英语口语考试问答题评分中提供了实证基础。除效果分析外,还探讨了三种方法在大规模口语考试实际应用中的优劣势,并对多模态大模型未来在口语评分任务上的应用潜力进行了展望。展开更多
Achieving robust walking for different stairs is one of the most challenging tasks for quadruped robots in real world.Traditional model-based methods heavily rely on environmental factors,are burdened by intricate mod...Achieving robust walking for different stairs is one of the most challenging tasks for quadruped robots in real world.Traditional model-based methods heavily rely on environmental factors,are burdened by intricate modelling complexities,and lack generalizability.The potential for advancements in adaptive locomotion control,often impeded by complex modelling processes,can be substantially enhanced through the application of Reinforcement Learning(RL).In this paper,a learning-based method is proposed to directionally enhance the stair-climbing skill of quadruped robots under different stair conditions.First,the general policy model based on proprioceptive perception is trained as a pre-training model.Then,the pre-training model was initialized,and different terrain information from the stairs was introduced for customized training to enhance the stair-climbing skill without affecting the existing locomotion performance.Finally,the customized control policy is deployed to the real robot to realize motion control in real environments.The experimental results demonstrate that the customized control policy can significantly improve the motion performance of quadruped robots when facing complex stair terrains and has certain generalizability in other complex terrains.The proposed algorithm can be extended to various terrestrial environments.展开更多
文摘针对于工业场所中密集场景下安全帽佩戴出现的漏检情况以及提高检测精度,提出了一种基于改进YOLOv5s的工业安全帽检测算法.首先,采用CIOU优化Soft-NMS对密集人群重叠的情况,减少重叠目标的漏检,从而提高了安全帽检测的准确性.其次,在网络的中间层添加辅助训练头引入丰富的梯度信息,最后,辅助训练头结合Optimal Transport Assignment添加到Loss中,通过最优的目标匹配,减少模型的漏检和误检的情况,从而提升模型的准确率和召回率.实验结果表明,改进后的算法平均精确值(mAP@50-90)值为68.3%,相对于原YOLOv5s算法提升了3.8%,准确率为92.3%,相较于原YOLOv5s算法提高了0.7%.
文摘The development of chemical technologies,which involves a multistage process covering laboratory research,scale‐up to industrial deployment,and necessitates interdisciplinary collaboration,is often accompanied by substantial time and economic costs.To address these challenges,in this work,we report ChemELLM,a domain‐specific large language model(LLM)with 70 billion parameters for chemical engineering.ChemELLM demonstrates state‐of‐the‐art performance across critical tasks ranging from foundational understanding to professional problem‐solving.It outperforms mainstream LLMs(e.g.,O1‐Preview,GPT‐4o,and DeepSeek‐R1)on ChemEBench,the first multidimensional benchmark for chemical engineering,which encompasses 15 dimensions across 101 distinct essential tasks.To support robust model development,we curated ChemEData,a purpose‐built dataset containing 19 billion tokens for pre‐training and 1 billion tokens for fine‐tuning.This work establishes a new paradigm for artificial intelligence‐driven innovation,bridging the gap between laboratory‐scale innovation and industrial‐scale implementation,thus accelerating technological advancement in chemical engineering.ChemELLM is publicly available at https://chemindustry.iflytek.com/chat.
文摘针对英语口语考试中常见的封闭类和开放类问答题,在中考口语模拟考试数据集合上,对比了基于特征工程、双向编码器小尺寸文本预训练模型(BERT)和单向解码器大尺寸文本预训练模型(Large Language Model,LLM)三种自动评分方法。实验结果表明,不管是在封闭类还是开放类问答题型上,基于LLM的自动评分方法表现最优,BERT模型次之,特征工程方法最差。在开放类问答上,基于LLM和BERT的自动评分相对特征工程方案优势更为明显,证明LLM和BERT模型在问答题这种需要语义理解的任务上具备天然优势,为大模型技术应用到大规模英语口语考试问答题评分中提供了实证基础。除效果分析外,还探讨了三种方法在大规模口语考试实际应用中的优劣势,并对多模态大模型未来在口语评分任务上的应用潜力进行了展望。
文摘Achieving robust walking for different stairs is one of the most challenging tasks for quadruped robots in real world.Traditional model-based methods heavily rely on environmental factors,are burdened by intricate modelling complexities,and lack generalizability.The potential for advancements in adaptive locomotion control,often impeded by complex modelling processes,can be substantially enhanced through the application of Reinforcement Learning(RL).In this paper,a learning-based method is proposed to directionally enhance the stair-climbing skill of quadruped robots under different stair conditions.First,the general policy model based on proprioceptive perception is trained as a pre-training model.Then,the pre-training model was initialized,and different terrain information from the stairs was introduced for customized training to enhance the stair-climbing skill without affecting the existing locomotion performance.Finally,the customized control policy is deployed to the real robot to realize motion control in real environments.The experimental results demonstrate that the customized control policy can significantly improve the motion performance of quadruped robots when facing complex stair terrains and has certain generalizability in other complex terrains.The proposed algorithm can be extended to various terrestrial environments.
