Unmanned aerial vehicle (UAV) experiments, multiple datasets from ground-based stations and satellite remote sensing platforms, and backward trajectory models were combined to investigate the characteristics and influential mechanisms of the air pollution episode that occurred in Nanjing during 3–4 December 2017. Before the experiments, the position of the detector mounted on a UAV that was minimally disturbed by the rotation of the rotors was analyzed based on computational fluid dynamics (CFD) simulations. The combined analysis indicated that the surface meteorological conditions—high relative humidity, low wind speed, and low temperature—were conducive to the accumulation of PM_2.5. Strongly intense temperature inversion layers and the low thickness of the atmospheric mixed layer could have resulted in elevated PM_2.5 mass concentrations. In the early stage, air pollution was affected by the synoptic circulation of the homogenous pressure field and low wind speeds, and the pollutants mainly originated from emissions from surrounding areas. The aggravated pollution was mainly attributed to the cold front and strong northwesterly winds above 850 hPa, and the pollutants mostly originated from the long-distance transport of emissions with northwesterly winds, mainly from the Beijing‒Tianjin‒Hebei (BTH) region and its surrounding areas. This long-distance transport predominated during this event. The air pollution level and aerosol optical depth (AOD) were positively correlated with respect to their spatial distributions; they could reflect shifts in areas of serious pollution. Pollution was concentrated in Anhui Province when it was alleviated in Nanjing. Polluted dust, polluted continental and smoke aerosols were primarily observed during this process. In particular, polluted dust aerosols accounted for a major part of the transport stage, and existed between the surface and 4 km. Moreover, the average extinction coefficient at lower altitudes (<1 km) was higher for aerosol deposition.

In order to make multirotor unmanned aerial vehicles (UAV) compose a desired dense formation and improve the practicality of UAV formation, a distributed algorithm based on fuzzy logic was proposed. The airflow created by multirotor UAVs was analyzed according to the structure of the multirotor UAV and the characteristic equation of the fluid. This paper presented a dynamic model for the process of formation of and path search algorithm based on this model. The membership function in this model combines the factors of position, flow field, and movement. Integrating the dynamic model and its desired position in formations, each UAV evaluates the surrounding points and then selects the direction for step motion. Through simulation, this algorithm was improved by a by-step formation approach, and the effectiveness of this method in dense formation of multirotor UAVs was proved.

Breakthroughs in natural language processing (NLP) by large-scale language models (LLMs) have led to superior performance in multilingual tasks such as translation, summarization, and Q&A. However, the size and complexity of these models raise challenges in terms of computational requirements, memory usage, and energy consumption. Quantization strategies, as a type of model compression technique, have gained attention for their advantages in reducing model size and accelerating inference speed. In this paper, we review the rapid development of quantization techniques for LLMs, and systematically explore methods such as post-training quantization (PTQ), quantization-aware fine-tuning (QAF), and quantization-aware training (QAT), which provide a comprehensive solution to the resource-intensive problem of LLMs from training to post-deployment. We also analyze state-of-the-art benchmarks and datasets to evaluate the effectiveness of quantitative methods in terms of performance retention and computational efficiency. The purpose of this review is to provide researchers with a snapshot of the latest progress in LLM quantization techniques, helping them to quickly grasp the dynamics of the field and understand the key techniques and challenges, so that they can more efficiently devote themselves to this evolving research area.

As Deep Neural Networks (DNNs) continue to increase in complexity, the computational demands of their training have become a significant bottleneck. Low-precision training has emerged as a crucial strategy, wherein full-precision values are quantized to lower precisions, reducing computational overhead while aiming to maintain model accuracy. While prior research has primarily focused on minimizing quantization noise and optimizing performance for specific models and tasks, a comprehensive understanding of the general principles governing low-precision computations across diverse DNN architectures has been lacking. In this paper, we address this gap by systematically analyzing the factors that influence low-precision matrix computations, which are fundamental to DNN training. We investigate three critical factors—accumulation in matrix calculations, the frequency of element usage, and the depth of matrices within the model—and their impact on low-precision training. Through controlled experiments on standard models, as well as customized experiments designed to isolate individual factors, we derive several key insights: layers with higher accumulation and matrices with lower usage frequencies demonstrate greater tolerance to low-precision noise, without significantly compromising the stability of model training. Additionally, while the depth of matrices influences the stability of matrix operations to some extent, it does not have a noticeable effect on the overall training outcomes. Our findings contribute to the development of generalizable principles for low-precision training, offering a systematic framework applicable across various DNN architectures. We provide empirical evidence supporting the strategic allocation of training bit-widths based on the analyzed factors, thereby enhancing the efficiency and effectiveness of DNN training in resource-constrained environments.

Low-precision training has emerged as a practical approach, saving the cost of time, memory, and energy during deep neural networks (DNNs) training. Typically, the use of lower precision introduces quantization errors that need to be minimized to maintain model performance, often neglecting to consider the potential benefits of reducing training precision. This paper rethinks low-precision training, highlighting the potential benefits of lowering precision: (1) low precision can serve as a form of regularization in DNN training by constraining excessive variance in the model; (2) layer-wise low precision can be seen as an alternative dimension of sparsity, orthogonal to pruning, contributing to improved generalization in DNNs. Based on these analyses, we propose a simple yet powerful technique–DPC (Decreasing Precision with layer Capacity), which directly assigns different bit-widths to model layers, without the need for an exhaustive analysis of the training process or any delicate low-precision criteria. Thorough extensive experiments on five datasets and fourteen models across various applications consistently demonstrate the effectiveness of the proposed DPC technique in saving computational cost (−16.21%–−44.37%) while achieving comparable or even superior accuracy (up to +0.68%, +0.21% on average). Furthermore, we offer feature embedding visualizations and conduct further analysis with experiments to investigate the underlying mechanisms behind DPC’s effectiveness, enhancing our understanding of low-precision training. Our source code will be released upon paper acceptance.