3D Gaussian Splatting (3D GS) has gained popularity due to its faster rendering speed and high-quality novel view synthesis. Some researchers have explored using 3D GS for reconstructing driving scenes. However, these methods often rely on various types of data, such as depth maps, 3D bounding boxes, and trajectories of moving objects. Additionally, the lack of annotations for synthesized images limits their direct application in downstream tasks. To address these issues, we propose EGSRAL, a 3D GS-based method that relies solely on training images without extra annotations. EGSRAL enhances 3D GS's capability to model both dynamic objects and static backgrounds and introduces a novel adaptor for auto labeling, generating corresponding annotations based on existing annotations. We also propose a grouping strategy for vanilla 3D GS to address perspective issues in rendering large-scale, complex scenes. Our method achieves state-of-the-art performance on multiple datasets without any extra annotation. For example, the PSNR metric reaches 29.04 on the nuScenes dataset. Moreover, our automated labeling can significantly improve the performance of 2D/3D detection tasks.

Traditional channel-wise pruning methods by reducing network channels struggle to effectively prune efficient CNN models with depth-wise convolutional layers and certain efficient modules, such as popular inverted residual blocks. Prior depth pruning methods by reducing network depths are not suitable for pruning some efficient models due to the existence of some normalization layers. Moreover, finetuning subnet with directly removing activation layers would corrupt the original model weights, hindering the pruned model from achieving high performance. To address these issues, we propose a novel depth pruning method for efficient models. Our approach proposes a novel block pruning strategy and progressive training method for the subnet. Additionally, we extend our pruning method to vision transformer models. Experimental results demonstrate that our method consistently outperforms existing depth pruning methods across various pruning configurations. We obtained three pruned ConvNeXtV1 models with our method applying on ConvNeXtV1, which surpass most SOTA efficient models with comparable inference performance. Our method also achieves state-of-the-art pruning performance on the vision transformer model.

Diffusion models have achieved remarkable progress in the field of image generation due to their outstanding capabilities. However, these models require substantial computing resources because of the multi-step denoising process during inference. While traditional pruning methods have been employed to optimize these models, the retraining process necessitates large-scale training datasets and extensive computational costs to maintain generalization ability, making it neither convenient nor efficient. Recent studies attempt to utilize the similarity of features across adjacent denoising stages to reduce computational costs through simple and static strategies. However, these strategies cannot fully harness the potential of the similar feature patterns across adjacent timesteps. In this work, we propose a novel pruning method that derives an efficient diffusion model via a more intelligent and differentiable pruner. At the core of our approach is casting the model pruning process into a SubNet search process. Specifically, we first introduce a SuperNet based on standard diffusion via adding some backup connections built upon the similar features. We then construct a plugin pruner network and design optimization losses to identify redundant computation. Finally, our method can identify an optimal SubNet through few-step gradient optimization and a simple post-processing procedure. We conduct extensive experiments on various diffusion models including Stable Diffusion series and DiTs. Our DiP-GO approach achieves 4.4 x speedup for SD-1.5 without any loss of accuracy, significantly outperforming the previous state-of-the-art methods.

This paper studies the problem of wide working conditions operational performance optimization of steam turbine units in coal-fired thermal power plants. A method based on big data association rule mining is proposed to establish relationships between a set of key controllable operating parameters and heat consumption rate of the units and then to determine optimal target values of the operating parameters. To deal with sparse data discretized from continuous unit operating data, a new association rule mining algorithm is proposed firstly. It utilizes a binary matrix to store data and introduces a search technique which combines linked list with pointers to mine frequent itemsets. These greatly eliminate the impact of sparsity of data i.e., large differences among transactions and scattered item distribution, on performance of the mining algorithm. Furthermore, in order to process large-scale data efficiently, a parallel implementation of the algorithm on Apache Spark platform is given. Finally, taking a steam turbine in a 600MW subcritical thermal power unit in China as an example, an overall optimization procedure including association rule mining and target value acquiring is presented. The mining results show that the target values calculated by the proposed optimization method can reduce heat consumption rate of the unit and improve economic benefits of the power plant significantly.Note to Practitioners—This research is motivated by the urgent requirement for energy saving and pollutant emission reducing in todays coal-fired power plants, especially in old thermal power plants. To achieve this objective, an effective method is to determine target values (also known as optimal values or benchmark values) of some key controllable operating parameters to realize operating optimization of the thermal power units. Existing target value determination approaches usually adopt the design values or the thermodynamic calculation values, which cannot truly represent the changes of unit operating characteristics. This research proposes a new big data-driven steam turbine unit target value determination method using association rule mining to reduce heat consumption rate of the unit. We not only design a new association rule mining algorithm but also give a parallel implementation of it on advanced big data distributed parallel computing framework Apache Spark, making the proposed optimization method efficiently mine large scale operating data. The proposed operating optimization method is based on the actual operating data of the unit. It has high reliability and good general applicability for different types of coal-fired power units. The optimization experiment on a steam turbine unit in China shows that the obtained target values are valuable for guiding optimization operation of the unit under wide working conditions. In the future, we will consider explicit pollutant emission in the optimization index and address multi-index operating optimization of the steam turbine unit.