检索结果-内蒙古大学图书馆

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Lei Yang Ming Sun Zhongwei Hu Zenan Zhang Wenxuan Yuan College of Electronic Information and Automation Civil Aviation University of China Tianjin China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Conventional statistical-driven synthetic aperture radar (SAR) imaging algorithms can only encode a single and/or static prior, leading to that limited features can be accessed quantitatively. To this end, a novel multi-task learning framework is proposed by devising a divide-and-conquer variational Bayesian (DC-VB) inference, so that elaborated features of interests can be exploited for a high-resolution SAR imagery. Specifically, a flexible generalized Gaussian distribution (GGD) and a customized hybrid probability distribution are introduced and employed for the priors of features of interests. To resolve the resultant complicated Bayesian inference, splitted random variables are incorporated, so that the joint posterior problem can be decomposed into multiple local problems that are easy to be solved. Simultaneously, dual variables are established for residual errors of the decomposition. To guarantee a global solution for the image of the target of interests, the data augmentation is employed to coordinate multiple local solutions. Therefore, the intended Bayesian inference works in a divide-and-conquer manner, which is superior in quantization of multiple features of high-resolution SAR imagery. It is capable of incorporating multiple priors in a fully-statistical probability and guaranteeing closed-form solutions of posterior distributions. Unavoidable propagation errors can be minimized in the DC-VB process. Raw SAR data is applied to validate the effectiveness of the proposed algorithm. Comparisons with conventions show the superiority in terms of qualitative and quantitative aspects.

关键词： Quantization (signal) Closed-form solutions Signal processing algorithms Multitasking Data augmentation Radar polarimetry inference algorithms Bayes methods Speech processing Synthetic aperture radar

来源：评论

学校读者我要写书评

暂无评论

Diffusion-Derep: Removing Sample Replication During Diffusion inference

Diffusion-Derep: Removing Sample Replication During Diffusio...

引用

IEEE International Symposium on Biomedical Imaging

作者： Nithin Gopalakrishnan Nair Vibashan Vishnukuamr Sharmini Vishal M. Patel Johns Hopkins University Baltimore MD USA

ISBN: (数字)9798331520526

ISBN: (纸本)9798331520533

Recent advancements in research on diffusion-based generative models have yielded excellent results in image, video, and 3D generation. Several open-source models for real-world images have recently been released, generating photorealistic results. However, in the medical domain, most major foundation diffusion models are held privately due to the data replication issue inherent in diffusion models. The data replication issue refers to a sampling problem observed in diffusion models, where the model tends to generate samples closely resembling the training data. This phenomenon becomes pronounced when a diffusion model is trained on a small dataset containing only a few hundred images. In this work, we address the data replication issue in diffusion models by proposing a novel inference process guidance algorithm. Specifically, during the inference process, we steer the generated samples away from the training samples by utilizing a similarity loss that quantifies the distance between the generated sample and the embedded representations of the training examples. To the best of our knowledge, this is the first attempt to tackle the sample replication problem in diffusion models. Our method can be easily integrated with existing methods, enhancing their performance in generating new, novel samples. We present results across 8 datasets across different domains and show that utilizing our sampling algorithm significantly reduces data mitigation and observed in our FID and SSCD experiments. Code: https://***/Vibashan/MedDataset

关键词： Training Solid modeling Three-dimensional displays Prevention and mitigation Training data Diffusion models Market research inference algorithms Tuning Biomedical imaging

来源：评论

学校读者我要写书评

暂无评论

Multi Modal English Vocabulary inference Algorithm Integrating ELECTRA and Knowledge Graph

Multi Modal English Vocabulary Inference Algorithm Integrati...

引用

Cybersecurity, Internet of Things and Soft Computing (CITSC), Asia-Europe Conference on

作者： Qinyue Wu Technical College for the Deaf Tianjin University of Technology Tianjin China

ISBN: (数字)9798331504205

ISBN: (纸本)9798331504212

In current English learning, the improvement of vocabulary inference ability is crucial to learners” language application ability. However, the existing single-mode algorithms are often unable to fully tap the deep relationship between words, resulting in insufficient accuracy of inference. This article aims to propose a multimodal English vocabulary inference algorithm that integrates ELECTRA model and knowledge graph to improve the accuracy and efficiency of vocabulary inference. Firstly, the ELECTRA model is used for deep learning of contextual semantics, and then a knowledge graph based vocabulary relationship network is constructed to integrate the semantic and relational information of vocabulary. Next, by integrating the contextual representation of ELECTRA with the node features of the knowledge graph through multimodal feature fusion technology, an inference model is constructed for training and optimization. The experimental results show that the algorithm achieves an accuracy of 87.5% in vocabulary inference tasks, which is more than 8% higher than traditional methods, and also improves inference efficiency by 25%. The multimodal inference algorithm that integrates ELECTRA and knowledge graph effectively solves the problem of information silos in vocabulary inference, providing English learners with more accurate and efficient vocabulary learning tools.

关键词： Deep learning Training Vocabulary Accuracy Semantics Knowledge graphs inference algorithms Internet of Things Optimization Context modeling

来源：评论

学校读者我要写书评

暂无评论

Online Network inference from Graph-Stationary Signals with Hidden Nodes

Online Network Inference from Graph-Stationary Signals with ...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Andrei Buciulea Madeline Navarro Samuel Rey Santiago Segarra Antonio G. Marques Dept. of Signal Theory and Communications Rey Juan Carlos University Madrid Spain Dept. of Electrical and Computer Engineering Rice University Houston TX USA

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Graph learning is the fundamental task of estimating unknown graph connectivity from available data. Typical approaches assume that not only is all information available simultaneously but also that all nodes can be observed. However, in many real-world scenarios, data can neither be known completely nor obtained all at once. We present a novel method for online graph estimation that accounts for the presence of hidden nodes. We consider signals that are stationary on the underlying graph, which provides a model for the unknown connections to hidden nodes. We then formulate a convex optimization problem for graph learning from streaming, incomplete graph signals. We solve the proposed problem through an efficient proximal gradient algorithm that can run in real-time as data arrives sequentially. Additionally, we provide theoretical conditions under which our online algorithm is similar to batch-wise solutions. Through experimental results on synthetic and real-world data, we demonstrate the viability of our approach for online graph learning in the presence of missing observations.

关键词： Heuristic algorithms Signal processing algorithms Estimation Signal processing Real-time systems inference algorithms Convex functions Acoustics Speech processing Optimization

来源：评论

学校读者我要写书评

暂无评论

algorithms for inference in specialized maps for recovering 3D hand pose

Algorithms for inference in specialized maps for recovering ...

引用

5th IEEE International Conference on Automatic Face and Gesture Recognition

作者： Rosales, R Sclaroff, S Boston Univ Dept Comp Sci Boston MA 02215 USA

ISBN: (纸本)0769516025

Learning in the Specialized Mappings Architecture (SMA) was presented for recovering 3D hand pose from visual features, however inference was not fully addressed. In this paper we tackle this aspect of the SMA model more thoroughly, and propose two inference algorithms, one deterministic and one probabilistic. SMA consists of several specialized forward (input to output) mapping functions that are estimated automatically from data and a known feedback or inverse function (that maps outputs to inputs). This allows the use of alternative conditional independence assumptions in the same model (derived from a forward and a feedback model) during learning and inference. The two proposed inference approximations, the mean output (MO) and the multiple sampling (MS) algorithms, are tested in recovering 3D hand pose from single images.

关键词： inference algorithms Computer science Output feedback Image sampling System testing Machine vision Biological system modeling Humans Object detection Skin

来源：评论

学校读者我要写书评

暂无评论

algorithms for the inference of Causality in Dynamic Processes: Application to Cardiovascular and Cerebrovascular Variability 37

Algorithms for the Inference of Causality in Dynamic Process...

引用

37th Annual International Conference of the IEEE-Engineering-in-Medicine-and-Biology-Society (EMBC)

作者： Faes, Luca Porta, Alberto Nollo, Giandomenico Univ Trento IRCS FBK Dept Ind Engn Trento Italy Univ Trento BIOtech Dept Ind Engn Trento Italy Univ Milan Dept Biomed Sci Hlth Milan Italy IRCCS Policlin San Donato Dept Cardiothorac Vasc Anesthesia & Intens Care Milan Italy

ISBN: (纸本)9781424492701

This study faces the problem of causal inference in multivariate dynamic processes, with specific regard to the detection of instantaneous and time-lagged directed interactions. We point out the limitations of the traditional Granger causality analysis, showing that it leads to false detection of causality when instantaneous and time-lagged effects coexist in the process structure. Then, we propose an improved algorithm for causal inference that combines the Granger framework with the approach proposed by Pearl for the study of causality among multiple random variables. This new approach is compared with the traditional one in theoretical and simulated examples of interacting processes, showing its ability to retrieve the correct structure of instantaneous and time-lagged interactions. These approaches for causal inference are then tested on the physiological variability series of heart period, arterial pressure and cerebral blood flow variability obtained in subjects with postural-related syncope during a tilt-test protocol.

关键词： blood blood vessels brain cardiovascular system haemodynamics random processes Pearl method arterial pressure cardiovascular variability causality inference algorithms cerebral blood flow variability cerebrovascular variability heart period instantaneous directed interactions multiple random variables multivariate dynamic processes physiological variability series postural-related syncope tilt-test protocol time-lagged directed interactions traditional Granger causality analysis Algorithm design and analysis Blood flow Heuristic algorithms inference algorithms Physiology Random variables Time series analysis

来源：评论

学校读者我要写书评

暂无评论

algorithms and complexity results for #SAT and Bayesian inference 44

Algorithms and complexity results for #SAT and Bayesian infe...

引用

44th Annual IEEE Symposium on Foundations of Computer Science

作者： Bacchus, F Dalmao, S Pitassi, T Univ Toronto Dept Comp Sci Toronto ON M5S 3G4 Canada

ISBN: (纸本)0769520405

Bayesian inference is an important problem with numerous applications in probabilistic reasoning. Counting satisfying assignments is a closely related problem of fundamental theoretical importance. In this paper we show that plain old DPLL equipped with memoization (an algorithm we call #DPLLCache) can solve both of these problems with time complexity that is at least as good as state-of-the-art exact algorithms, and that it cat, also achieve the best known time-space tradeoff. We then proceed to show that there are instances where #DPLLCache can achieve an exponential speedup over existing algorithms.

关键词： Application software Bayesian methods Computational modeling Computer science Computer simulation Government inference algorithms Polynomials

来源：评论

学校读者我要写书评

暂无评论

Causal Discovery for Rolling Bearing Fault Under Missing Data: From the Perspective of Causal Effect and Information Flow

引用

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT 2025年 74卷

作者： Ding, Xu Wu, Hao Wang, Junlong Xu, Juan Xin, Miao Hefei Univ Technol Sch Mech Engn Anhui Prov Key Lab Aerosp Struct Parts Forming Tec Hefei 230002 Peoples R China Hefei Univ Technol Inst Ind & Equipment Technol Hefei 230002 Peoples R China Hefei Univ Technol Sch Comp & Informat Hefei 230061 Peoples R China State Key Lab High End Compressor & Syst Technol Hefei 230031 Anhui Peoples R China Chinese Acad Sci Inst Automat Beijing 100190 Peoples R China

Missing data phenomenon is ubiquitous in many domains which imposes difficulties on the reconstruction of causal relationships due to incomplete datasets. Currently, although causal discovery techniques tackling data missing complete at random (MCAR) and missing at random (MAR) are known to the public, attaining causal structure from data missing not at random (MNAR) is still beyond our reach. In this article, a causal effect and information flow Peter Clark (CEIFPC) algorithm is proposed, which uses a missingness graph to represent the missing mechanism and combines causal effect correction criteria (CECC) with information flow to recover the potential causal structure. First, in the sketchy stage, a test-wise deletion Peter Clark (TD-PC) algorithm for causal skeleton learning is introduced. Second, in the refining stage, by employing CECC, one can identify erroneous edges within the causal skeleton and subsequently eliminate them. Finally, in the orientation stage, the directional strategies from the Peter Clark (PC) algorithm and information flow are incorporated to accurately determine the directions of causal edges. Our proposed CEIFPC algorithm not only theoretically yields asymptotically correct results but also demonstrates superior performance in recovering the potential causal structure under nonrandom missing data. As shown by experimental results, our proposed method, CEIFPC, exhibits good performance on both synthetic and real bearing failure datasets.

关键词： Skeleton Mars Rolling bearings Refining Fault diagnosis Data models Accuracy Machine learning inference algorithms Bayes methods Causal discovery causal effect information flow missing data rolling bearing

来源：评论

学校读者我要写书评

暂无评论

An Automated Method to Correct Artifacts in Neural Text-to-Speech Models

IEEE Transactions on Audio, Speech and Language Processing

引用

IEEE Transactions on Audio, Speech and Language Processing 2025年 33卷 1375-1388页

作者： Seongyeop Jeong June Sig Sung Inchul Hwang Jaesik Choi Kim Jaechul Graduate School of Artificial Intelligence KAIST Daejeon South Korea Ineeji Seongnam South Korea Mobile eXperience Business Samsung Electronics Suwon South Korea

Recent advances in deep learning technology have enabled high-quality speech synthesis, and text-to-speech models are widely used in a variety of applications. However, even state-of-the-art models still produce artificial speech, highlighting the need to correct errors in synthesized speech. Traditional speech correction methodologies face a number of challenges, such as the inefficiency of manually specifying errors, the need to retrain models, and the need for additional data to correct synthesized speech errors. In this paper, we present a novel approach that detects and corrects contextual errors within a model to improve synthetic speech without requiring additional resources or model retraining. Specifically, we analyze the inherent limitations of neural network encoders responsible for contextualizing input sentences and propose a method that automatically identifies abnormal encoder context vectors. We also introduce a correction algorithm that enhances the quality of speech by correcting the incorrect relationships between phonemes that cause abnormal encoder context vectors. Our algorithm demonstrated a 25.86% and 4.69% reduction in alignment errors and a 2.25% and 0.63% improvement in objective metrics such as Fréchet Wav2Vec distance for the Tacotron2 and VITS models. In addition, our algorithm improved the comparative mean opinion scores, a subjective evaluation, by 1.34 and 0.52. The results support the feasibility of a novel approach to identify, correct, and improve defects automatically in neural speech synthesis models. The results support the feasibility of a novel approach to identify, correct, and improve defects automatically in neural speech synthesis models.

关键词： Vectors Context modeling Data models Training Analytical models Neurons Speech enhancement inference algorithms Biological neural networks Text to speech

来源：评论

学校读者我要写书评

暂无评论

DeepSeek: Paradigm Shifts and Technical Evolution in Large AI Models

引用

IEEE/CAA Journal of Automatica Sinica 2025年第5期12卷 841-858页

作者： Luolin Xiong Haofen Wang Xi Chen Lu Sheng Yun Xiong Jingping Liu Yanghua Xiao Huajun Chen Qing-Long Han Yang Tang Key Laboratory of Smart Manufacturing in Energy Chemical Process Ministry of Education and the Engineering Research Center of Process System Engineering Ministry of Education East China University of Science and Technology Shanghai China College of Design and Innovation Tongji University Shanghai China Shanghai Key Laboratory of Data Science School of Computer Science Fudan University Shanghai China School of Software Beihang University Beijing China School of Information Science and Engineering East China University of Science and Technology Shanghai China AZFT Joint Laboratory for Knowledge Engine Hangzhou Innovation Center College of Computer Science and Technology Zhejiang University Hangzhou China School of Science Computing and Engineering Technologies Swinburne University of Technology Melbourne VIC Australia

DeepSeek, a Chinese artificial intelligence (AI) startup, has released their V3 and R1 series models, which attracted global attention due to their low cost, high performance, and open-source advantages. This paper begins by reviewing the evolution of large AI models focusing on paradigm shifts, the mainstream large language model (LLM) paradigm, and the DeepSeek paradigm. Subsequently, the paper highlights novel algorithms introduced by DeepSeek, including multi-head latent attention (MLA), mixture-of-experts (MoE), multi-token prediction (MTP), and group relative policy optimization (GRPO). The paper then explores DeepSeek's engineering breakthroughs in LLM scaling, training, inference, and system-level optimization architecture. Moreover, the impact of DeepSeek models on the competitive AI landscape is analyzed, comparing them to mainstream LLMs across various fields. Finally, the paper reflects on the insights gained from DeepSeek's innovations and discusses future trends in the technical and engineering development of large AI models, particularly in data, training, and reasoning.

关键词： Training Technological innovation Foundation models Large language models Reinforcement learning Prediction algorithms Market research Cognition inference algorithms Optimization

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：