Deep neural networks(DNNs)have achieved great success in many data processing applications.However,high computational complexity and storage cost make deep learning difficult to be used on resource-constrained devices...Deep neural networks(DNNs)have achieved great success in many data processing applications.However,high computational complexity and storage cost make deep learning difficult to be used on resource-constrained devices,and it is not environmental-friendly with much power cost.In this paper,we focus on low-rank optimization for efficient deep learning techniques.In the space domain,DNNs are compressed by low rank approximation of the network parameters,which directly reduces the storage requirement with a smaller number of network parameters.In the time domain,the network parameters can be trained in a few subspaces,which enables efficient training for fast convergence.The model compression in the spatial domain is summarized into three categories as pre-train,pre-set,and compression-aware methods,respectively.With a series of integrable techniques discussed,such as sparse pruning,quantization,and entropy coding,we can ensemble them in an integration framework with lower computational complexity and storage.In addition to summary of recent technical advances,we have two findings for motivating future works.One is that the effective rank,derived from the Shannon entropy of the normalized singular values,outperforms other conventional sparse measures such as the?_1 norm for network compression.The other is a spatial and temporal balance for tensorized neural networks.For accelerating the training of tensorized neural networks,it is crucial to leverage redundancy for both model compression and subspace training.展开更多
目的探讨基于MRI的椎体骨质量评分(vertebral bone quality score,VBQ)和终板骨质量评分(endplate bone quality score,EBQ)在经椎间孔腰椎椎间融合(transforaminal lumbar interbody fusion,TLIF)术后cage沉降中的预测价值。方法因腰...目的探讨基于MRI的椎体骨质量评分(vertebral bone quality score,VBQ)和终板骨质量评分(endplate bone quality score,EBQ)在经椎间孔腰椎椎间融合(transforaminal lumbar interbody fusion,TLIF)术后cage沉降中的预测价值。方法因腰椎退行性疾病在我院行TLIF手术的226例患者,根据术后有无cage沉降将患者分为沉降组和非沉降组,比较两组患者VBQ和EBQ评分。通过多元回归分析cage沉降的危险因素,并根据受试者工作特征曲线下面积(AUC)评估VBQ和EBQ预测TLIF术后cage沉降的能力。结果226例患者中30例出现术后cage沉降。沉降组VBQ(3.8±0.4)分,EBQ(5.1±0.7)分,明显高于非沉降组(3.1±0.6)分和(4.2±1.0)分,差异有统计学意义(P<0.001)。多元回归分析显示VBQ(OR=4.258,95%CI:1.983~9.142,P<0.001)和EBQ(OR=1.971,95%CI:1.212~3.203,P=0.006)评分越高,发生cage沉降风险也越大。受试者工作特征曲线结果显示VBQ的AUC为0.843,EBQ的AUC是0.864。VBQ和EBQ预测cage沉降的最佳阈值分别为3.480(敏感性90%;特异性75.5%)和4.620(敏感性96.7%;特异性74.5%)。结论术前VBQ或EBQ评分越高,TLIF术后发生cage沉降风险越大。其中EBQ可能是一个更好的预测融合术后cage沉降的指标。展开更多
鉴于码率控制在视频编码应用中的重要性,综述了针对新一代国际视频编码标准HEVC(high efficiency video coding)的码率控制技术研究进展。介绍了码率控制的基本工作原理,通过调节编码器参数使输出码率符合信道带宽限制或满足存储设备的...鉴于码率控制在视频编码应用中的重要性,综述了针对新一代国际视频编码标准HEVC(high efficiency video coding)的码率控制技术研究进展。介绍了码率控制的基本工作原理,通过调节编码器参数使输出码率符合信道带宽限制或满足存储设备的容量需求,主要包含比特分配和比特控制2个部分。然后,介绍了HEVC开发过程中接收的2个重要码率控制技术提案,从码率模型和比特分配两方面阐述了提案中的URQ(unified rate-quantization)模型码率控制算法和R-λ模型码率控制算法。依据比特控制中的码率模型选择,把近年出现的HEVC码率控制算法分为Q域或λ域方法,并分别展开论述和技术点归纳;对HEVC扩展及其相应的码率控制进行了简要阐述。最后,结合码率控制仍需解决的问题,提出一些值得进一步探索的工作。展开更多
基金supported by the National Natural Science Foundation of China(62171088,U19A2052,62020106011)the Medico-Engineering Cooperation Funds from University of Electronic Science and Technology of China(ZYGX2021YGLH215,ZYGX2022YGRH005)。
文摘Deep neural networks(DNNs)have achieved great success in many data processing applications.However,high computational complexity and storage cost make deep learning difficult to be used on resource-constrained devices,and it is not environmental-friendly with much power cost.In this paper,we focus on low-rank optimization for efficient deep learning techniques.In the space domain,DNNs are compressed by low rank approximation of the network parameters,which directly reduces the storage requirement with a smaller number of network parameters.In the time domain,the network parameters can be trained in a few subspaces,which enables efficient training for fast convergence.The model compression in the spatial domain is summarized into three categories as pre-train,pre-set,and compression-aware methods,respectively.With a series of integrable techniques discussed,such as sparse pruning,quantization,and entropy coding,we can ensemble them in an integration framework with lower computational complexity and storage.In addition to summary of recent technical advances,we have two findings for motivating future works.One is that the effective rank,derived from the Shannon entropy of the normalized singular values,outperforms other conventional sparse measures such as the?_1 norm for network compression.The other is a spatial and temporal balance for tensorized neural networks.For accelerating the training of tensorized neural networks,it is crucial to leverage redundancy for both model compression and subspace training.
文摘目的探讨基于MRI的椎体骨质量评分(vertebral bone quality score,VBQ)和终板骨质量评分(endplate bone quality score,EBQ)在经椎间孔腰椎椎间融合(transforaminal lumbar interbody fusion,TLIF)术后cage沉降中的预测价值。方法因腰椎退行性疾病在我院行TLIF手术的226例患者,根据术后有无cage沉降将患者分为沉降组和非沉降组,比较两组患者VBQ和EBQ评分。通过多元回归分析cage沉降的危险因素,并根据受试者工作特征曲线下面积(AUC)评估VBQ和EBQ预测TLIF术后cage沉降的能力。结果226例患者中30例出现术后cage沉降。沉降组VBQ(3.8±0.4)分,EBQ(5.1±0.7)分,明显高于非沉降组(3.1±0.6)分和(4.2±1.0)分,差异有统计学意义(P<0.001)。多元回归分析显示VBQ(OR=4.258,95%CI:1.983~9.142,P<0.001)和EBQ(OR=1.971,95%CI:1.212~3.203,P=0.006)评分越高,发生cage沉降风险也越大。受试者工作特征曲线结果显示VBQ的AUC为0.843,EBQ的AUC是0.864。VBQ和EBQ预测cage沉降的最佳阈值分别为3.480(敏感性90%;特异性75.5%)和4.620(敏感性96.7%;特异性74.5%)。结论术前VBQ或EBQ评分越高,TLIF术后发生cage沉降风险越大。其中EBQ可能是一个更好的预测融合术后cage沉降的指标。
文摘鉴于码率控制在视频编码应用中的重要性,综述了针对新一代国际视频编码标准HEVC(high efficiency video coding)的码率控制技术研究进展。介绍了码率控制的基本工作原理,通过调节编码器参数使输出码率符合信道带宽限制或满足存储设备的容量需求,主要包含比特分配和比特控制2个部分。然后,介绍了HEVC开发过程中接收的2个重要码率控制技术提案,从码率模型和比特分配两方面阐述了提案中的URQ(unified rate-quantization)模型码率控制算法和R-λ模型码率控制算法。依据比特控制中的码率模型选择,把近年出现的HEVC码率控制算法分为Q域或λ域方法,并分别展开论述和技术点归纳;对HEVC扩展及其相应的码率控制进行了简要阐述。最后,结合码率控制仍需解决的问题,提出一些值得进一步探索的工作。
