In the field of simultaneous localization and mapping (SLAM), visual odometry (VO) always has great application prospects. In recent years, with the progress in the field of machine learning, methods based on neural n...
详细信息
Optical thermometry based on the upconversion(UC)luminescence intensity ratio(LIR)has attracted considerable attention because of its feasibility for achievement of accurate non-contact temperature *** with traditiona...
详细信息
Optical thermometry based on the upconversion(UC)luminescence intensity ratio(LIR)has attracted considerable attention because of its feasibility for achievement of accurate non-contact temperature *** with traditional UC phosphors,optical thermometry based on UC single crystals can achieve faster response and higher sensitivity due to the stability and high thermal conductivity of the single *** this study,a high-quality 5 at%Yb^(3+)and 1 at%Ho^(3+)co-doped Gd_(0.74)Y_(0.2)TaO_(4)single crystal was grown by the Czochralski(Cz)method,and the structure of the as-grown crystal was ***,the UC luminescent properties and optical thermometry behaviors of this crystal were *** 980 nm wavelength excitation,green and red UC luminescence lines at 550 and 650 nm and corresponding to the^(5)F_(4)/^(5)S_(2)→^(5)I_(8)and^(5)F_(5)→^(5)I_(8)transitions of Ho^(3+),respectively,were *** green and red UC emissions involved a two-photon mechanism,as evidenced by the analysis of power-dependent UC emission *** temperature-dependent UC emission spectra were measured in the temperature range of 330–660 K to assess the optical temperature sensing *** 660 K,the maximum relative sensing sensitivity(S_(r))was determined to be 0.0037 K^(−1).These results highlight the signifcant potential of Yb,Ho:GYTO single crystal for optical temperature sensors.
Federated reinforcement learning (FRL) uses data from multiple partners interacting with the environment to train a global decision model while maintaining data privacy. In specific situations, it is necessary to prot...
详细信息
The increasing radio frequency interference(RFI)is a well-recognized problem in radio astronomy *** and Fast Radio Bursts(FRBs)are high-priority science targets of the ongoing Commercial Radio Astronomy FAST Survey(CR...
详细信息
The increasing radio frequency interference(RFI)is a well-recognized problem in radio astronomy *** and Fast Radio Bursts(FRBs)are high-priority science targets of the ongoing Commercial Radio Astronomy FAST Survey(CRAFTS).To improve the quality of RFI removal in searches of pulsars and FRBs based on CRAFTS multi-beam data,we here propose an intuitive but powerful RFI mitigation pipeline(CCF-ST).The“CCF-ST”is a spatial filter constructed by signal cross-correlation function(CCF)and Sum-Threshold(ST)*** RFI marking result is saved in a“mask”file,a binary format for RFI masks in *** known pulsars,PSR B0525-21,PSR B0621-04,and PSR J0943+2252 from CRAFTS L-band 19 beams data are used for evaluation of the performance of CCF-ST in comparison with other methods,such as PRESTO’s“rfifind”,ArPLS-ST and *** result shows that CCF-ST can reduce effective data loss rate and improves the detected signal-to-noise ratio of the pulsations by~26%and~18%respectively compared with PRESTO’s“rfifind”and *** CCF-ST also has the advantage of low computational cost,e.g.,reducing the time consumption by~40%and memory consumption by~90%compared with *** expect that the new RFI mitigation and analysis toolkit(CCF-ST)demonstrated in this paper can be applied to CRAFTS and other multi-beam telescope observations to improve the data quality and efficiency of pulsar and FRB searches.
The uncapacitated facility location problem (UFLP) is a well-known combinational optimization problem, attracting numerous heuristic and meta-heuristic methods. However, these effective algorithms still encounter chal...
详细信息
The paper investigates the robustness and parallel scaling properties of a novel physical factorization preconditioner with algebraic multigrid subsolves in the iterative solution of a cell-centered finite volume disc...
详细信息
The paper investigates the robustness and parallel scaling properties of a novel physical factorization preconditioner with algebraic multigrid subsolves in the iterative solution of a cell-centered finite volume discretization of the threedimensional multi-group radiation diffusion *** key idea is to take advantage of a particular kind of block factorization of the resulting system matrix and approximate the left-hand block matrix selectively spurred by parallel processing *** spectral property of the preconditioned matrix is then *** practical strategy is considered sequentially and in ***,numerical results illustrate the numerical robustness,computational efficiency and parallel strong and weak scalabilities over the real-world structured and unstructured coupled problems,showing its competitiveness with many existing block preconditioners.
EEG-based fatigue driving monitoring has important application value in road traffic safety, and the ultimate goal of the research is the development and use of wearable devices, and too many EEG channels in practical...
详细信息
Big data drive multidimensional convergence and profound innovations among industries and provide novel ways of exploring the world. As they significantly create economic and social value, big data meaningfully impact...
详细信息
Big data drive multidimensional convergence and profound innovations among industries and provide novel ways of exploring the world. As they significantly create economic and social value, big data meaningfully impact the implementation and management of information security and privacy *** technologies are used to protect the security and entire life cycle of big data. The demand for this technology is multiplied when the data are stored in the cloud. They are stored in the cloud in the form of ciphertext, and the driving requirement for data retrieval, sharing, and manipulation places the security of data at risk. The all-or-nothing approach of traditional cryptography systems cannot realize the release and processing of data information with flexible and increasingly fine granularity. Consequently, dealing with the relationship between privacy protection and data utilization, as well as navigating the blurry boundaries between subverting either plaintext or ciphertext, has become a research focus of the current cryptographic trend for protecting big data security. Presently, there are many studies designed to solve these ***, security requirements and source encryption mode for future big data systems and applications are elaborated. Then, focusing on the practical security and functionality of the big data life cycle, including storage, retrieval, sharing, calculation, statistical analysis, and utilization, the research being conducted based on those functions is reviewed. For each cryptographic technology that meets the requirement of each type of big data security or application, security and efficiency comments and sufficient comparison analyses of cryptography schemes or protocols are provided; moreover, the current general problems and development trends are expounded. Because the current innovative research on cryptographic technology was primarily based on the development or improvement of a single solution, the study on
In this paper,we present a comprehensive overview of artificial intelligence(AI)computing systems for large language models(LLMs)*** rapid advancement of LLMs in recent years,coupled with the widespread adoption of al...
详细信息
In this paper,we present a comprehensive overview of artificial intelligence(AI)computing systems for large language models(LLMs)*** rapid advancement of LLMs in recent years,coupled with the widespread adoption of algorithms and applications such as BERT,ChatGPT,and DeepSeek,has sparked significant interest in this *** classify LLMs into encoder-only,encoder-decoder,and decoder-only models,and briefly analyze their training and infer-ence processes to emphasize their substantial need for computational *** operations depend heavily on AI-specific accelerators like GPUs(graphics processing units),TPUs(tensor processing units),and MLUs(machine learning units).However,as the gap widens between the increasing complexity of LLMs and the current capabilities of accelerators,it becomes essential to adopt heterogeneous computing systems optimized for distributed environments to manage the growing computational and memory requirements of *** delve into the execution and scheduling of LLM algo-rithms,underlining the critical role of distributed computing strategies,memory management enhancements,and boosting computational *** paper clarifies the complex relationship between algorithm design,hardware infrastructure,and software optimization,and provides an in-depth understanding of both the software and hardware infrastructure sup-porting LLMs training,offering insights into the challenges and potential avenues for future development and deployment.
Traditional clustering algorithms often focus on the most fine-grained information and achieve clustering by calculating the distance between each pair of data points or implementing other calculations based on points...
详细信息
暂无评论