检索结果-内蒙古大学图书馆

Feature-wise and Sample-wise Adaptive Transfer Learning for High-dimensional Linear Regression

学校读者我要写书评

暂无评论

arXiv 2024年

作者： He, Zelin Sun, Ying Liu, Jingyuan Li, Runze Department of Statistics Pennsylvania State University United States School of Electrical Engineering and Computer Science Pennsylvania State University United States MOE Key Laboratory of Econometrics Department of Statistics and Data Science School of Economics Wang Yanan Institute for Studies in Economics Fujian Key Lab of Statistics Xiamen University China

We consider the transfer learning problem in the high dimensional linear regression setting, where the feature dimension is larger than the sample size. To learn transferable information, which may vary across features or the source samples, we propose an adaptive transfer learning method that can detect and aggregate the feature-wise (F-AdaTrans) or sample-wise (S-AdaTrans) transferable structures. We achieve this by employing a fused-penalty, coupled with weights that can adapt according to the transferable structure. To choose the weight, we propose a theoretically informed, data-driven procedure, enabling F-AdaTrans to selectively fuse the transferable signals with the target while filtering out non-transferable signals, and S-AdaTrans to obtain the optimal combination of information transferred from each source sample. We show that, with appropriately chosen weights, F-AdaTrans achieves a convergence rate close to that of an oracle estimator with known transferable structure, and S-AdaTrans recovers existing near-minimax optimal rates as a special case. The effectiveness of the proposed method is validated using both synthetic and real data, demonstrating favorable performance compared to the existing methods. Copyright © 2024, The Authors. All rights reserved.

关键词： Transfer learning

Improving Nonlocal Means Method via a No-Reference Image Content Metric for MRI Denoising

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2025年第4期23卷 735-741页

作者： Xin Hou Jianwu Li Yao Lu Zhengchao Dong Beijing Key Laboratory of Intelligent Information Technology School of Computer Science and Technology Beijing Institute of Technology Beijing China Brain Imaging Lab Columbia University New York NY United States New York State Psychiatric Institute New York NY United States

The patch matching of the traditional Non-local means (NLM) filter mainly depends on structure similarity and cannot adapt to the patch rotation or mirroring transformation. Therefore, designing a measure with rotationally invariant similarity is of significant importance for improving the effectiveness of patch comparison of NLM. This paper proposes to apply a no-reference image content metric with the rotation-invariance to NLM for denoising Magnetic resonance (MR) images. The metric measures quantitatively the content of a patch in an image, including sharpness, contrast, and geometric features such as textures and edges. The metric values for every patch are computed and added into the Gaussian matching kernel of NLM so as to effectively perform patch matching. The main advantage of the proposed method is that it does not need to rotate patches in different orientations during patch matching. Experimental results show that the proposed method is superior to the traditional NLM, the state-of-the-art method Block-matching and 3-D (BM3D) filtering and the Unbiased NLM (UNLM) for MRI denoising.

关键词： Measurement Matched filters Three-dimensional displays Magnetic resonance imaging Noise reduction Superresolution Magnetic resonance Kernel Rotation measurement Image denoising

A Hybrid RAG System with Comprehensive Enhancement on Complex Reasoning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Yuan, Ye Liu, Chengwu Yuan, Jingyang Sun, Gongbo Li, Siqi Zhang, Ming State Key Laboratory of Multimedia Information Processing School of Computer Science PKU-Anker Embodied AI Lab Peking University Beijing100871 China School of Computer Data & Information Sciences University of Wisconsin-Madison MadisonWI United States

Retrieval-augmented generation (RAG) is a framework enabling large language models (LLMs) to enhance their accuracy and reduce hallucinations by integrating external knowledge bases. In this paper, we introduce a hybrid RAG system enhanced through a comprehensive suite of optimizations that significantly improve retrieval quality, augment reasoning capabilities, and refine numerical computation ability. We refined the text chunks and tables in web pages, added attribute predictors to reduce hallucinations, conducted LLM Knowledge Extractor and Knowledge Graph Extractor, and finally built a reasoning strategy with all the references. We evaluated our system on the CRAG dataset through the Meta CRAG KDD Cup 2024 Competition. Both the local and online evaluations demonstrate that our system significantly enhances complex reasoning capabilities. In local evaluations, we have significantly improved accuracy and reduced error rates compared to the baseline model, achieving a notable increase in scores. In the meanwhile, we have attained outstanding results in online assessments, demonstrating the performance and generalization capabilities of the proposed system. The source code for our system is released in https://***/shizueyy/crag-new. Copyright © 2024, The Authors. All rights reserved.

关键词： Knowledge graph

Spin-UP: Spin Light for Natural Light Uncalibrated Photometric Stereo

学校读者我要写书评

暂无评论

Spin-UP: Spin Light for Natural Light Uncalibrated Photometr...

Conference on computer Vision and Pattern Recognition (CVPR)

作者： Zongrui Li Zhan Lu Haojie Yan Boxin Shi Gang Pan Qian Zheng Xudong Jiang Rapid-Rich Object Search (ROSE) Lab Interdisciplinary Graduate Programme Nanyang Technological University School of Electrical and Electronic Engineering Nanyang Technological University The State Key Lab of Brain-Machine Intelligence Zhejiang University College of Computer Science and Technology Zhejiang University National Key Laboratory for Multimedia Information Processing School of Computer Science Peking University National Engineering Research Center of Visual Technology School of Computer Science Peking University

ISBN: (数字)9798350353006

ISBN: (纸本)9798350353013

Natural Light Uncalibrated Photometric Stereo (NaUPS) relieves the strict environment and light assumptions in classical Uncalibrated Photometric Stereo (UPS) methods. However, due to the intrinsic ill-posedness and high-dimensional ambiguities, addressing NaUPS is still an open question. Existing works impose strong assumptions on the environment lights and objects' material, restricting the effectiveness in more general scenarios. Alternatively, some methods leverage supervised learning with intricate models while lacking interpretability, resulting in a biased estimation. In this work, we propose Spin Light Uncalibrated Photometric Stereo (Spin-UP), an unsupervised method to tackle NaUPS in various environment lights and objects. The proposed method uses a novel setup that captures the object's images on a rotatable platform, which mitigates NaUPS's ill-posedness by reducing unknowns and provides reliable priors to alleviate NaUPS's ambiguities. Leveraging neural inverse rendering and the proposed training strategies, Spin-UP recovers surface normals, environment light, and isotropic reflectance under complex natural light with low computational cost. Experiments have shown that Spin-UP outperforms other supervised / unsupervised NaUPS meth-ods and achieves state-of-the-art performance on synthetic and real-world datasets. Codes and data are available at https://***/LMozart/CVPR2024-SpinUP.

关键词： Reflectivity Training Shape Supervised learning Anisotropic Rendering (computer graphics) Structural engineering

TalkCLIP: Talking Head Generation with Text-Guided Expressive Speaking Styles

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Ma, Yifeng Wang, Suzhen Ding, Yu Ma, Bowen Lv, Tangjie Fan, Changjie Hu, Zhipeng Deng, Zhidong Yu, Xin Li, Lincheng Department of Computer Science and Technology BNRist THUAI State Key Laboratory of Intelligent Technology and Systems Tsinghua University China Fuxi AI Lab Netease Zhejiang Hangzhou China The School of Computer Science The University of Queensland Brisbane Australia

Audio-driven talking head generation has drawn growing attention. To produce talking head videos with desired facial expressions, previous methods rely on extra reference videos to provide expression information, which may be difficult to find and hence limits their usage. In this work, we propose TalkCLIP, a framework that can generate talking heads where the expressions are specified by natural language, hence allowing for specifying expressions more conveniently. To model the mapping from text to expressions, we first construct a text-video paired talking head dataset where each video has diverse text descriptions that depict both coarse-grained emotions and fine-grained facial movements. Leveraging the proposed dataset, we introduce a CLIP-based style encoder that projects natural language-based descriptions to the representations of expressions. TalkCLIP can even infer expressions for descriptions unseen during training. TalkCLIP can also use text to modulate expression intensity and edit expressions. Extensive experiments demonstrate that TalkCLIP achieves the advanced capability of generating photo-realistic talking heads with vivid facial expressions guided by text descriptions. Copyright © 2023, The Authors. All rights reserved.

关键词： Natural language processing systems

Backdoor Defense via Deconfounded Representation Learning

学校读者我要写书评

暂无评论

Backdoor Defense via Deconfounded Representation Learning

Conference on computer Vision and Pattern Recognition (CVPR)

作者： Zaixi Zhang Qi Liu Zhicai Wang Zepu Lu Qingyong Hu Anhui Province Key Lab of Big Data Analysis and Application School of Computer Science and Technology University of Science and Technology of China State Key Laboratory of Cognitive Intelligence Hefei Anhui China University of Science and Technology of China Hong Kong University of Science and Technology

Deep neural networks (DNNs) are recently shown to be vulnerable to backdoor attacks, where attackers embed hidden backdoors in the DNN model by injecting a few poisoned examples into the training dataset. While extensive efforts have been made to detect and remove backdoors from backdoored DNNs, it is still not clear whether a backdoor-free clean model can be directly obtained from poisoned datasets. In this paper, we first construct a causal graph to model the generation process of poisoned data and find that the backdoor attack acts as the confounder, which brings spurious associations between the input images and target labels, making the model predictions less reliable. Inspired by the causal understanding, we propose the Causality-inspired Backdoor Defense (CBD), to learn deconfounded representations for reliable classification. Specifically, a backdoored model is intentionally trained to capture the confounding effects. The other clean model dedicates to capturing the desired causal effects by minimizing the mutual information with the confounding representations from the backdoored model and employing a sample-wise re-weighting scheme. Extensive experiments on multiple benchmark datasets against 6 state-of-the-art attacks verify that our proposed defense method is effective in reducing backdoor threats while maintaining high accuracy in predicting benign samples. Further analysis shows that CBD can also resist potential adaptive attacks. The code is available at https://***/zaixizhang/CBD.

关键词：

SHaRPose: Sparse High-Resolution Representation for Human Pose Estimation

学校读者我要写书评

暂无评论

arXiv 2023年

作者： An, Xiaoqi Zhao, Lin Gong, Chen Wang, Nannan Wang, Di Yang, Jian PCA Lab Key Lab of Intelligent Perception and Systems for High-Dimensional Information of Ministry of Education Jiangsu Key Lab of Image and Video Understanding for Social Security School of Computer Science and Engineering Nanjing University of Science and Technology China State Key Laboratory of Integrated Services Networks Xidian University China

High-resolution representation is essential for achieving good performance in human pose estimation models. To obtain such features, existing works utilize high-resolution input images or fine-grained image tokens. However, this dense high-resolution representation brings a significant computational burden. In this paper, we address the following question: "Only sparse human keypoint locations are detected for human pose estimation, is it really necessary to describe the whole image in a dense, high-resolution manner?" Based on dynamic transformer models, we propose a framework that only uses Sparse High-resolution Representations for human Pose estimation (SHaRPose). In detail, SHaRPose consists of two stages. At the coarse stage, the relations between image regions and keypoints are dynamically mined while a coarse estimation is generated. Then, a quality predictor is applied to decide whether the coarse estimation results should be refined. At the fine stage, SHaRPose builds sparse high-resolution representations only on the regions related to the keypoints and provides refined high-precision human pose estimations. Extensive experiments demonstrate the outstanding performance of the proposed method. Specifically, compared to the state-of-the-art method ViTPose, our model SHaRPose-Base achieves 77.4 AP (+0.5 AP) on the COCO validation set and 76.7 AP (+0.5 AP) on the COCO test-dev set, and infers at a speed of 1.4× faster than ViTPose-Base. Code is available at https://***/AnxQ/sharpose. Copyright © 2023, The Authors. All rights reserved.

关键词： Machine learning

KnowLogic: A Benchmark for Commonsense Reasoning via Knowledge-Driven Data Synthesis

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Zhan, Weidong Wang, Yue Hu, Nan Xiao, Liming Ma, Jingyuan Qin, Yuhang Li, Zheng Yang, Yixin Deng, Sirui Ding, Jinkun Ma, Wenhan Li, Rui Luo, Weilin Liu, Qun Sui, Zhifang Center for Chinese Linguistics Department of Chinese Language and Literature Peking University China School of Computer Science State Key Laboratory of Multimedia Information Processing Peking University China Huawei Noah’s Ark Lab China

Current evaluations of commonsense reasoning in LLMs are hindered by the scarcity of natural language corpora with structured annotations for reasoning tasks. To address this, we introduce KnowLogic, a benchmark generated through a knowledge-driven synthetic data strategy. KnowLogic integrates diverse commonsense knowledge, plausible scenarios, and various types of logical reasoning. One of the key advantages of KnowLogic is its adjustable difficulty levels, allowing for flexible control over question complexity. It also includes fine-grained labels for in-depth evaluation of LLMs’ reasoning abilities across multiple dimensions. Our benchmark consists of 3,000 bilingual (Chinese and English) questions across various domains, and presents significant challenges for current LLMs, with the highest-performing model achieving only 69.57%. Our analysis highlights common errors, such as misunderstandings of low-frequency commonsense, logical inconsistencies, and overthinking. This approach, along with our benchmark, provides a valuable tool for assessing and enhancing LLMs’ commonsense reasoning capabilities and can be applied to a wide range of knowledge domains. Our data and code can be found at https://***/pokerwf/ KnowLogic. Copyright © 2025, The Authors. All rights reserved.

关键词： Natural language processing systems

Survey on depth and RGB image-based 3D hand shape and pose estimation

学校读者我要写书评

暂无评论

Virtual Reality & Intelligent Hardware 2021年第3期3卷 207-234页

作者： Lin HUANG Boshen ZHANG Zhilin GUO Yang XIAO Zhiguo CAO Junsong YUAN Department of Computer Science and Engineering State University of New York at BuffaloNY 14260USA YouTu Lab TencentShanghai 201101China Department of Computer Science Fu Foundation School of Engineering and Applied ScienceColumbia UniversityNY 10027USA National Key Laboratory of Science and Technology on Multispectral Information Processing School of Artificial Intelligence and AutomationHuazhong University of Science and TechnologyWuhan 430074China

The field of vision-based human hand three-dimensional(3D)shape and pose estimation has attracted significant attention recently owing to its key role in various applications,such as natural human computer *** the availability of large-scale annotated hand datasets and the rapid developments of deep neural networks(DNNs),numerous DNN-based data-driven methods have been proposed for accurate and rapid hand shape and pose ***,the existence of complicated hand articulation,depth and scale ambiguities,occlusions,and finger similarity remain *** this study,we present a comprehensive survey of state-of-the-art 3D hand shape and pose estimation approaches using RGB-D *** RGB-D cameras,hand datasets,and a performance analysis are also discussed to provide a holistic view of recent *** also discuss the research potential of this rapidly growing field.

关键词： Hand survey 3D hand pose estimation Hand shape reconstruction Hand-object interactions RGB-D cameras