检索结果-内蒙古大学图书馆

Enhancing offline reinforcement learning for wastewater treatment via transition filter and prioritized approximation loss

引用

NEUROCOMPUTING 2025年 636卷

作者： Yang, Ruyue Wang, Ding Li, Menghua Cui, Chengyu Qiao, Junfei Beijing Univ Technol Sch Informat Sci & Technol Beijing Peoples R China Beijing Univ Technol Beijing Key Lab Computat Intelligence & Intelligen Beijing Peoples R China Beijing Univ Technol Beijing Lab Smart Environm Protect Beijing Peoples R China Beijing Univ Technol Beijing Inst Artificial Intelligence Beijing Peoples R China State Grid Corp China State Grid Beijing Chaoyang Power Supply Branch Beijing Peoples R China

Wastewater treatment plays a crucial role in urban society, requiring efficient control strategies to optimize its performance. In this paper, we propose an enhanced offline reinforcement learning (RL) approach for wastewater treatment. Our algorithm improves the learning process. It uses a transition filter to sort out low- performance transitions and employs prioritized approximation loss to achieve prioritized experience replay with uniformly sampled loss. Additionally, the variational autoencoder is introduced to address the problem of distribution shift in offline RL. The proposed approach is evaluated on a nonlinear system and wastewater treatment simulation platform, demonstrating its effectiveness in achieving optimal control. The contributions of this paper include the development of an improved offline RL algorithm for wastewater treatment and the integration of transition filtering and prioritized approximation loss. Evaluation results demonstrate that the proposed algorithm achieves lower tracking error and cost.

关键词： Offline reinforcement learning Wastewater treatment variational autoencoder Adaptive dynamic programming

来源：评论

学校读者我要写书评

暂无评论

Ground-motion generations using Multi-label Conditional Embedding-conditional Denoising Diffusion Probabilistic Model (ML-cDDPM)

引用

SOIL DYNAMICS AND EARTHQUAKE ENGINEERING 2025年 191卷

作者： Huang, Yanwen Sun, Xiaodan You, Junjie Yang, Cheng Lu, Dagang Zhang, Jie Southwest Jiaotong Univ Dept Civil Engn Chengdu 610093 Peoples R China Natl Engn Res Ctr Technol Geol Disaster Prevent La Chengdu 610031 Peoples R China Harbin Inst Technol Dept Civil Engn Harbin 150090 Peoples R China Tongji Univ Key Lab Geotech & Underground Engn Minist Educ Shanghai 200092 Peoples R China Tongji Univ Dept Geotech Engn Shanghai 200092 Peoples R China

Current studies on ground motion prediction equations require the categorization of earthquake records by regions with different attenuation patterns. Conversely, machine learning based on entire datasets provides an opportunity to explore the factors that dominate ground motion at a site and their complex interactions. In this study, Multi-label Conditional Embedding is proposed to modify a conditional Denoising Diffusion Probabilistic Model (cDDPM) and achieve ground motion prediction of earthquake scenarios. A database was created for neural network training, consisting of 7154 horizontal ground motion records from 105 earthquakes selected from the Next Generation Attenuation (NGA)-West2 database of the Pacific Earthquake Engineering Research Center (PEER). Each record was labeled using four conditional parameters: VS30, F, MW, and Rrup. To embed multi-label and location information into the neural network, One-hot Encoding and Positional Encoding techniques were integrated. Hence, a Multi-label Conditional Embedding-conditional Denoising Diffusion Probabilistic Model (ML-cDDPM) was constructed for ground motion prediction. The model was used to simulate ground motions of past earthquakes and was compared with recorded motions. The model was also compared with two other neural-network-based prediction models. The comparisons demonstrate the reliability of ML-cDDPM in simulating ground motions for earthquake scenarios and its superiority over the other two models in representing the complexity of ground motion.

关键词： Ground motion generation Denoising diffusion probabilistic model Earthquake scenarios Multi-label conditional embedding variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

Developing an Urban Landscape Fumigation Service Robot: A Machine-Learned, Gen-AI-Based Design Trade Study

引用

APPLIED SCIENCES-BASEL 2025年第4期15卷 2061-2061页

作者： Chittoor, Prithvi Krishna Dandumahanti, Bhanu Priya Veerajagadheswar, Prabakaran Samarakoon, S. M. Bhagya P. Muthugala, M. A. Viraj J. Elara, Mohan Rajesh Singapore Univ Technol & Design Engn Prod Dev Pillar Singapore 487372 Singapore

Generative AI (Gen-AI) revolutionizes design by leveraging machine learning to generate innovative solutions. It analyzes data to identify patterns, creates tailored designs, enhances creativity, and allows designers to explore complex possibilities for diverse industries. This study uses a Gen-AI design generation process to develop an urban landscape fumigation service robot. This study proposes a machine-learned multimodal and feedback-based variational autoencoder (MMF-VAE) model that incorporates a readily available spraying robot dataset and includes design considerations from various research efforts to ensure real-time deployability. The objective is to demonstrate the effectiveness of data-driven and feedback-based approaches in generating design specifications for a fumigation robot with the targeted requirements of autonomous navigation, precision spraying, and an extended runtime. The design generation process comprises three stages: (1) parameter fixation, emphasizing functionality-based and aesthetic-based specifications;(2) design specification generation using the proposed MMF-VAE model with and without a spraying robot dataset;and (3) robot development based on the generated specifications. A comparative analysis evaluated the impact of the dataset-driven design generation. The design generated with the dataset proved more feasible and optimized for real-world deployment with the integration of multimodal inputs and iterative feedback refinement. A real-time prototype was then constructed using the model's parametric constraints and tested in actual fumigation scenarios to validate operational viability. This study highlights the transformative potential of Gen-AI in robotic design workflows.

关键词： generative design variational autoencoder multimodal design fumigation robot urban landscape

来源：评论

学校读者我要写书评

暂无评论

Study of Hybrid Machine Learning Multiaxial Low-Cycle Fatigue Life Prediction Model of CP-Ti

引用

FATIGUE & FRACTURE OF ENGINEERING MATERIALS & STRUCTURES 2025年第5期48卷 2309-2324页

作者： Ma, Tian-Hao Zhang, Wei Chang, Le Zhao, Jian-Ping Zhou, Chang-Yu Nanjing Tech Univ Sch Mech & Power Engn Nanjing Peoples R China

Symmetric and asymmetric multiaxial low-cycle fatigue tests were conducted on commercially pure titanium under different control modes and multiaxial strain/stress ratios to establish a reliable hybrid physics and data-driven method. Optimized analysis formula-based models are proposed to provide reliable physical information first. Based on the dataset enhanced by the nonlinear variational autoencoder method, a hybrid VAE-ANN model is established and trained, developed using the Pearson correlation coefficient analysis and Leaky ReLU activation function. Through a series of fatigue life prediction validations under both symmetric and asymmetric loading conditions, the VAE-ANN model demonstrates excellent prediction accuracy, broad generalization capability, and strong compatibility, achieving the lowest average absolute relative error of 6.76% under symmetric and 22.61% under asymmetric loading conditions.

关键词： artificial neural network commercial pure titanium life prediction multiaxial low-cycle fatigue variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

Machine learning assisted crystallographic reconstruction from atom probe tomographic images

引用

JOURNAL OF PHYSICS-CONDENSED MATTER 2025年第3期37卷 035901-035901页

作者： Pu, Jie-Ming Chen, Shuai Zhang, Tong-Yi Shanghai Univ Mat Genome Inst Shanghai 200444 Peoples R China Shanghai Univ Shanghai Frontier Sci Ctr Mechanoinformat Shanghai 200444 Peoples R China Hong Kong Univ Sci & Technol Guangzhou Guangzhou Municipal Key Lab Mat Informat Sustainab Adv Mat Thrust Guangzhou 511400 Guangdong Peoples R China

Atom probe tomography (APT) is a powerful technique for three-dimensional (3D) atomic-scale imaging, enabling the accurate analysis on the compositional distribution at the nanoscale. How to accurately reconstruct crystallographic information from APT data, however, is still a great challenge due to the intrinsic nature of the APT technique. In this paper, we propose a novel approach that consists of the modified forward simulation process and the backward machine learning process to recover the tested crystal from APT data. The high-throughput forward simulations on Al single crystals of different orientations generate 10 000 original 3D images and data augmentation is implemented on the original images, resulting in 100 000 3D images. The big data allows the development of deep learning models and three deep learning algorithms of Convolutional Neural Network (CNN), Vision Transformer (ViT), and variational autoencoder (VAE) are used in the backward process. After training, the ViT model performs superior than the CNN and VAE models, which can recover the crystalline orientation outstandingly, as evaluated by the coefficient of determination R-2 and the Mean Percent Error (MPE), viz., R-2 = 0.93 and MPE = 0.43%, R-2 = 0.97 and MPE = 0.35%, and R-2 = 0.93 and MPE = 0.77% for the rotation angles phi, psi and theta, respectively, on the test dataset. The present work clearly demonstrates the capability of deep learning models in the recovery of the tested crystals from APT data, thereby paving the way for the further development of large artificial intelligent models of APT.

关键词： atom probe tomography high-throughput simulation convolutional neural network vision of transformer variational autoencoder machine learning

来源：评论

学校读者我要写书评

暂无评论

An Efficient Speech Synthesizer: A Hybrid Monotonic Architecture for Text-to-speech via VAE & LPC-Net with Independent Sentence Length

引用

CIRCUITS SYSTEMS AND SIGNAL PROCESSING 2025年 1-25页

作者： Nallabala, Naveen Kumar Souprayen, Balamurugan Ramasamy, Maruthamuthu Penumarti, Seshu Kumar Navuluri, Nuthan Chandu Madanapalle Inst Technol & Sci Dept Comp Applicat Madanapalle Andhra Pradesh India JAIN Sch Comp Sci & Informat Technol Bengaluru Karnataka India Univ East London Dept Architecture Comp & Engn London England

This research delineates the development of a hybrid architecture for text-to-speech (TTS) synthesis, termed the Efficient Speech Synthesizer (ESS). ESS leverages an end-to-end training paradigm to holistically optimize all the parameters, thereby facilitating the synthesis of high-fidelity speech with remarkable efficiency, irrespective of sentence length. The architecture integrates state-of-the-art feed-forward methodologies, including variational autoencoders and linear predictive coding techniques, which impose monotonic constraints on sequence alignment while incurring negligible computational overhead. The primary objective of this study is to engineer a TTS model characterized by minimal complexity and latency, rendering it suitable for deployment across a broad spectrum of computational platforms, ranging from high-performance systems to resource-constrained IoT devices and smartphones.

关键词： Efficient Speech Synthesizer Non-autoregressive Autoregressive Monotonic alignment Linear predictive coding variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

Integrating physics of the problem into data-driven methods to enhance elastic full-waveform inversion with uncertainty quantification

引用

PHYSICA D-NONLINEAR PHENOMENA 2025年 472卷

作者： Negahdari, Vahid Moghadasi, Seyed Reza Razvan, Mohammad Reza Sharif Univ Technol Dept Math Sci Tehran 111559415 Iran

Full-Waveform Inversion (FWI) is a nonlinear iterative seismic imaging technique that, by reducing the misfit between recorded and predicted seismic waveforms, can produce detailed estimates of subsurface geophysical properties. Nevertheless, the strong nonlinearity of FWI can trap the optimization in local minima. This issue arises due to factors such as improper initial values, the absence of low frequencies in the measurements, noise, and other related considerations. To address this challenge and with the advent of advanced machine-learning techniques, data-driven methods, such as deep learning, have attracted significantly increasing attention in the geophysical community. Furthermore, the elastic wave equation should be included in FWI to represent elastic effects accurately. The intersection of data-driven techniques and elastic scattering theories presents opportunities and challenges. In this paper, by using the knowledge of elastic scattering (physics of the problem) and integrating it with machine learning techniques, we propose methods for the solution of time- harmonic FWI to enhance accuracy compared to pure data-driven and physics-based approaches. Moreover, to address uncertainty quantification, by modifying the structure of the variational autoencoder, we introduce a probabilistic deep learning method based on the physics of the problem that enables us to explore the uncertainties of the solution. According to the limited availability of datasets in this field and to assess the performance and accuracy of the proposed methods, we create a comprehensive dataset close to reality and conduct a comparative analysis of the presented approaches to it.

关键词： Full waveform inversion Inverse elastic scattering Data-driven technique Deep learning variational autoencoder Seismic imaging Uncertainty quantification

来源：评论

学校读者我要写书评

暂无评论

Machine sound anomaly detection based on dual-channel feature fusion variational auto-encoder

引用

APPLIED INTELLIGENCE 2025年第7期55卷 1-17页

作者： Zhang, Chen Wei, Yongkang Wang, Xiaofeng Wu, Xiaoxuan Zhu, Xuhui Hefei Univ Dept Artificial Intelligence & Big Data Hefei 230601 Anhui Peoples R China Hefei Univ Technol Intelligent Interconnected Syst Lab Anhui Prov Hefei 230009 Anhui Peoples R China

With the increasing intelligence and automation of industrial equipment, the technology for detecting equipment anomalies has become increasingly important. Compared to image-based anomaly detection methods, sound-based anomaly detection methods have the advantages of being non-intrusive, real-time and having lower data collection costs. These advantages make them highly valuable for research. Currently, deep learning methods that focus on spectrogram reconstruction have become widely utilized in the field of machine sound anomaly detection research. However, previous methods only attempted to mitigate the impact of noise without enabling the model to fully learn the distribution of sound features during the reconstruction process. In this paper, a novel Dual-Channel Feature Fusion variational autoencoder (DCFF-VAE) is proposed to effectively improve its reconstruction ability and help it better learn the normal sound features. In this method, the deep features extracted from the convolution layer and bidirectional gated cycle unit in the encoder are integrated by means of concatenation to make full use of the important features in the sound. Subsequently, grouped deconvolution is applied in the decoder to reduce model complexity while enhancing its perceptual ability for features. Additionally, during the anomaly detection phase, anomaly scores are calculated based on the Mahalanobis distance to better capture the differences between normal and abnormal sounds. Anomaly detection experiments conducted on five types of machines demonstrate that DCFF-VAE not only achieves the best stability but also surpasses the best comparison method by 3.14% and 1.21% in AUC and pAUC metrics, respectively.

关键词： Anomalous sound detection Unsupervised learning Feature fusion variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

A deep learning model for inter-fraction head and neck anatomical changes in proton therapy

引用

PHYSICS IN MEDICINE AND BIOLOGY 2025年第6期70卷 065011-065011页

作者： Burlacu, Tiberiu Hoogeman, Mischa Lathouwers, Danny Perko, Zoltan Delft Univ Technol Fac Appl Sci Delft Netherlands Univ Med Ctr Rotterdam Erasmus MC Canc Inst Dept Radiotherapy Rotterdam Netherlands HollandPTC Consortium Delft Netherlands

Objective. To assess the performance of a probabilistic deep learning based algorithm for predicting inter-fraction anatomical changes in head and neck patients. Approach. A probabilistic daily anatomy model (DAM) for head and neck patients DAM (DAMHN) is built on the variational autoencoder architecture. The model approximates the generative joint conditional probability distribution of the repeat computed tomography (rCT) images and their corresponding masks on the planning CT images (pCT) and their masks. The model outputs deformation vector fields, which are used to produce possible rCTs and associated masks. The dataset is composed of 93 patients (i.e. 315 pCT-rCT pairs), 9 (i.e. 27 pairs) of which were set aside for final testing. The performance of the model is assessed based on the reconstruction accuracy and the generative performance for the set aside patients. Main results. The model achieves a DICE score of 0.83 and an image similarity score normalized cross-correlation of 0.60 on the test set. The generated parotid glands, spinal cord and constrictor muscle volume change distributions and center of mass shift distributions were also assessed. For all organs, the medians of the distributions are close to the true ones, and the distributions are broad enough to encompass the real observed changes. Moreover, the generated images display anatomical changes in line with the literature reported ones, such as the medial shifts of the parotids glands. Significance. DAMHN is capable of generating realistic anatomies observed during the course of the treatment and has applications in anatomical robust optimization, treatment planning based on plan library approaches and robustness evaluation against inter-fractional changes.

关键词： proton therapy deep learning variational autoencoder anatomy changes

来源：评论

学校读者我要写书评

暂无评论

Glissando-Net: Deep Single View Category Level Pose Estimation and 3D Reconstruction

引用

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 2025年第4期47卷 2298-2312页

作者： Sun, Bo Kang, Hao Guan, Li Li, Haoxiang Mordohai, Philippos Hua, Gang Adobe Inc San Jose CA 95110 USA ByteDance Inc Bellevue WA USA Meta Real Labs Menlo Pk CA USA Pixocial Technol Bellevue WA USA Stevens Inst Technol Hoboken NJ USA Dolby Labs Bellevue WA USA

We present a deep learning model, dubbed Glissando-Net, to simultaneously estimate the pose and reconstruct the 3D shape of objects at the category level from a single RGB image. Previous works predominantly focused on either estimating poses (often at the instance level), or reconstructing shapes, but not both. Glissando-Net is composed of two auto-encoders that are jointly trained, one for RGB images and the other for point clouds. We embrace two key design choices in Glissando-Net to achieve a more accurate prediction of the 3D shape and pose of the object given a single RGB image as input. First, we augment the feature maps of the point cloud encoder and decoder with transformed feature maps from the image decoder, enabling effective 2D-3D interaction in both training and prediction. Second, we predict both the 3D shape and pose of the object in the decoder stage. This way, we better utilize the information in the 3D point clouds presented only in the training stage to train the network for more accurate prediction. We jointly train the two encoder-decoders for RGB and point cloud data to learn how to pass latent features to the point cloud decoder during inference. In testing, the encoder of the 3D point cloud is discarded. The design of Glissando-Net is inspired by codeSLAM. Unlike codeSLAM, which targets 3D reconstruction of scenes, we focus on pose estimation and shape reconstruction of objects, and directly predict the object pose and a pose invariant 3D reconstruction without the need of the code optimization step. Extensive experiments, involving both ablation studies and comparison with competing methods, demonstrate the efficacy of our proposed method, and compare favorably with the state-of-the-art.

关键词： Shape Three-dimensional displays Point cloud compression Solid modeling Training Decoding Pose estimation Image reconstruction Predictive models Transforms 3D shape reconstruction 3D pose estimation single view 3D shape estimation variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：