检索结果-内蒙古大学图书馆

11th International Conference on Learning Representations, ICLR 2023

作者： Liu, Zhenghao Xiong, Chenyan Lv, Yuanhuiyi Liu, Zhiyuan Yu, Ge Department of Computer Science and Technology Northeastern University Shenyang China Microsoft Research Redmond United States Department of Computer Science and Technology Tsinghua University Beijing China Institute for Artificial Intelligence Tsinghua University Beijing China State Key Lab on Intelligent Technology and Systems Tsinghua University Beijing China

This paper presents Universal Vision-Language Dense Retrieval (UniVL-DR), which builds a unified model for multi-modal retrieval. UniVL-DR encodes queries and multi-modality resources in an embedding space for searching candidates from different modalities. To learn a unified embedding space for multi-modal retrieval, UniVL-DR proposes two techniques: 1) Universal embedding optimization strategy, which contrastively optimizes the embedding space using the modality-balanced hard negatives;2) Image verbalization method, which bridges the modality gap between images and texts in the raw data space. UniVL-DR achieves the state-of-the-art on the multi-modal open-domain question answering benchmark, WebQA, and outperforms all retrieval models on the two subtasks, text-text retrieval and text-image retrieval. It demonstrates that universal multi-modal search is feasible to replace the divide-and-conquer pipeline with a united model and also benefits single/cross modality tasks. All source codes of this work are available at https://***/OpenMatch/UniVL-DR. © 2023 11th International Conference on Learning Representations, ICLR 2023. All rights reserved.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

The EGY-DRiVeS Dataset: Illuminating the Dynamics of Egyptian Road Traffic Behaviors

The EGY-DRiVeS Dataset: Illuminating the Dynamics of Egyptia...

引用

2024 International Conference on computer and Applications, ICCA 2024

作者： Saadawy, Zeyad H. Elfar, Mohamed A. Aly, Menna M. Adel, Momen Bahy, Mohamed A. Elias, Catherine M. C-DRiVeS Lab: Cognitive Driving Research in Vehicular Systems Cairo Egypt German University in Cairo Faculty of Media Engineering and Technology Computer Science and Engineering Department Egypt

ISBN: (纸本)9798350367560

This paper presents a comprehensive dataset of Egyptian roads integrated with chaotic scenarios captured by EGY-DRiVeS' lab golf car to aid in developing the autonomous driving experience. Using Velodyne 3D laser scanners, C922 Pro HD webcams, and iPhone 11 Pro cameras, a cumulative recording time of 3 hours and 48 minutes was documented. This included 1 hour and 15 minutes of LiDAR recordings, 1 hour and 45 minutes of camera recordings, and 48 minutes and 13 seconds of synced recordings. The dataset encompasses diverse scenarios and objects, ranging from urban environments with dynamic and static objects to rural highways. In-campus roads and intersections are also included, providing real-world scenarios such as vehicular and pedestrian behavior within campus premises. Each part of the dataset is annotated with 2D bounding boxes offering detailed information about vehicles, pedestrians and road objects. The provided data are calibrated, synchronized, and timestamped. Alongside the dataset, data format specifications, annotated data samples, and all supplementary recordings are provided upon request. © 2024 IEEE.

关键词： Network security

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Classification and Segmentation of Brain MRI Images

Deep Learning Classification and Segmentation of Brain MRI I...

引用

2023 International Conference on New Frontiers in Communication, Automation, Management and Security, ICCAMS 2023

作者： Pooja Ahuja, S. Likhith, S.R. Shiva Balan, A. Navya, K. Bengaluru India Bio-intelligence Lab Presidency University Department of School of Computer Science and Engineering and Information Science Bengaluru India Nagarjuna College of Engineering and Technology Department of Ise Bengaluru India

ISBN: (纸本)9798350317060

Unique Clinical pictures have a basic impact on the specialist's capability to the right conclusion and in the treatment of the patient. Diagnostic imaging can quickly identify lesions thanks to sophisticated algorithms, and it is essential to glean characteristics from photographs. In a number of studies, a number of algorithms have been incorporated into imaging in medicine. Using convolutional neural network, CNN's fundamental architecture is constructed using picture component extraction. To be able to circumvent the limitations imposed by both machine vision and human vision, the study is extended to include CNN with multiple input channels for visual feature extraction. This study uses approximately 3300 MRI samples from Kaggle to investigate four classifications: tumors of the pituitary gland, meningioma, glioma, and none. The implemented BrainNet has a accuracy of 98.31 percent and a validation accuracy of 87.80 percent. In order to determine which strategy performed better, deep architectures like InceptionNet and ResNet were also put through tests with or without learning through transfer. © 2023 IEEE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Enhancing Highway Safety with LSTM-based Vehicle Intention Prediction (HVIP)

Enhancing Highway Safety with LSTM-based Vehicle Intention P...

引用

2024 International Conference on computer and Applications, ICCA 2024

作者： Amer, Lujina Elias, Catherine M. C-DRiVeS Lab: Cognitive Driving Research in Vehicular Systems Cairo Egypt German University in Cairo Faculty of Media Engineering and Technology Computer Science and Engineering Department Egypt

ISBN: (纸本)9798350367560

Predicting vehicles' motion on highways has become crucial for enhancing road safety and traffic flow. Deep learning, which reached exceptional results in various applications, is now the leading approach for vehicle motion prediction. This work presents a deep learning-based model using Long Short-Term Memory (LSTM) networks to classify vehicle intentions into lane-keeping, left lane-changing, and right lane-changing. Utilizing the PREVENTION dataset, which provides naturalistic driving data, a sequence of centre points, longitudinal distances, lateral distances and yaw angles of a vehicle has been extracted to train the model to predict lane changes effectively. The model consists of three LSTM layers, a dense output layer and a drop out added between layers to prevent overfitting. Experiments focusing on optimizing the model parameters, learning rate, sequence length, and batch size, were conducted to determine their impact on prediction effectiveness. The best-performing model showed significant results in lane changing prediction, achieving an accuracy of 84%, precision of 89%, and recall of 82%. Future work will aim to train the model on more complex features, to account for vehicle inter-dependencies and interactions, and test model on various highway scenarios for a more reliable system. © 2024 IEEE.

关键词： Vehicle safety

来源：评论

学校读者我要写书评

暂无评论

MetaSTC: A Backbone Agnostic Spatio-Temporal Framework for Traffic Forecasting 24

MetaSTC: A Backbone Agnostic Spatio-Temporal Framework for T...

引用

24th IEEE International Conference on Data Mining, ICDM 2024

作者： Xu, Kexin Yu, Zhemeng Gao, Yucen Zhang, Songjian Fang, Jun Gao, Xiaofeng Chen, Guihai Shanghai Jiao Tong University MoE Key Lab of Artificial Intelligence Department of Computer Science and Engineering Shanghai China Didi Chuxing Technology Co. Beijing China

ISBN: (纸本)9798331506681

Traffic flow prediction is a critical issue in transportation engineering and presents distinct challenges when handling large-scale datasets in the real world. Existing complex spatio-temporal forecasting paradigms use the same parameters to fit traffic sequences with varying spatio-temporal features, and tend to train an average performance model over different time series. This approach greatly reduces their accuracy when applied to larger road networks. Moreover, the significant differences in traffic data distribution from one city to another can also pose great challenges. The same model may be excellent for one city and mediocre when applied to another. To this end, we propose a Meta Backbone Agnostic Spatio-Temporal Clustering Framework for Traffic Forecasting on Large-Scale Road Networks named MetaSTC. We tackle the disparities of spatio-temporal features of traffic flow through a spatio-temporal clustering-based strategy. We design meta-learner for large-scale road network that dynamically extracts the shared information across roads in the same sub-task. In this way, the model can represent task-specific details with a simpler model and make quick and accurate predictions. Our paradigm is backbone-agnostic and can be combined with different traffic prediction models, solving the problem caused by the difference in data distribution. Extensive experimental results conducted on real-world traffic dataset demonstrate the high accuracy and computational efficiency of our model over SOTA approaches. © 2024 IEEE.

关键词： Large datasets

来源：评论

学校读者我要写书评

暂无评论

SRender: Boosting Neural Radiance Field Efficiency via Sensitivity-Aware Dynamic Precision Rendering 57

SRender: Boosting Neural Radiance Field Efficiency via Sensi...

引用

57th Annual IEEE/ACM International Symposium on Microarchitecture, MICRO 2024

作者： Song, Zhuoran He, Houshu Liu, Fangxin Hao, Yifan Song, Xinkai Jiang, Li Liang, Xiaoyao Shanghai Jiao Tong University Department of Computer Science and Engineering Shanghai China Institute of Computing Technology Chinese Academy of Sciences State Key Lab of Processors Beijing China

ISBN: (纸本)9798350350579

Neural Radiance Field (NeRF) holds immense promise for generating photo-realistic images and videos. How-ever, the computation and memory demands significantly impede its applicability in real-Time tasks such as virtual reality and augmented reality. While accelerators dedicatedly designed for NeRF have emerged, they may not be the optimal solution as they only consider the optimization of full-precision NeRF models, leading to unsatisfactory memory and computation saving. This paper proposes Sensitivity-Aware Dynamic Precision Rendering, abbreviated as SRender, an algorithm-hardware co-design framework that borrows the idea of Adaptive Rendering commonly used in conventional graphics rendering to accelerate the rendering process of NeRF exploiting the fine-grained per-ray and per-point levels sensitivity. The fundamental idea behind SRender is that rays and points with high sensitivity require high-precision data types, while the insensitive rays and points do not, allowing them to be sacrificed for greater acceleration. Accordingly, the algorithm part of SRender dynamically configures the precision of the hash table and the multilayer perceptron (MLP), and the hardware part features a specialized accelerator that exploits such mixed-precision memory access and computation patterns. Specifically, we predict the sensitive rays and points based on low-precision hash encoding and feature computation. Then, we trigger a high-precision rendering again on the sensitive ones to guarantee the rendering quality. The hardware design of SRender includes a sensitivity prediction engine, coarse-Then-fine-grained encoding engines, and a recovery engine, which translates the computational saving in the algorithm to real speedup. Using the coarse-Then-fine-grained encoding engine, we maximize the data reuse on-chip, thereby reducing the stress of off-chip memory bandwidth. Experiments show that the edge-level SRender can achieve 3.94x and 34.06x speedup over state-of-The-Art NeRF

关键词： Encoding (symbols)

来源：评论

学校读者我要写书评

暂无评论

Geometry-Calibrated DRO: Combating Over-Pessimism with Free Energy Implications 41

Geometry-Calibrated DRO: Combating Over-Pessimism with Free ...

引用

41st International Conference on Machine Learning, ICML 2024

作者： Liu, Jiashuo Wu, Jiayun Wang, Tianyu Zou, Hao Li, Bo Cui, Peng Department of Computer Science and Technology Tsinghua University China Department of Industrial Engineering and Operations Research Columbia University United States Zhongguancun Lab China School of Economics and Management Tsinghua University China

Machine learning algorithms minimizing average risk are susceptible to distributional shifts. Distributionally Robust Optimization (DRO) addresses this issue by optimizing the worst-case risk within an uncertainty set. However, DRO suffers from over-pessimism, leading to low-confidence predictions, poor parameter estimations as well as poor generalization. In this work, we conduct a theoretical analysis of a probable root cause of over-pessimism: excessive focus on noisy samples. To alleviate the impact of noise, we incorporate data geometry into calibration terms in DRO, resulting in our novel Geometry-Calibrated DRO (GCDRO) for regression. We establish the connection between our risk objective and the Helmholtz free energy in statistical physics, and this free-energy-based risk can extend to standard DRO methods. Leveraging gradient flow in Wasserstein space, we develop an approximate minimax optimization algorithm with a bounded error ratio and elucidate how our approach mitigates noisy sample effects. Comprehensive experiments confirm GCDRO's superiority over conventional DRO methods. Copyright 2024 by the author(s)

关键词：

来源：评论

学校读者我要写书评

暂无评论

Transcriptomic and Macroscopic Architectures of Multimodal Covariance Network Reveal Molecular–Structural–Functional Co-alterations

引用

Research 2024年第1期2023卷 471-487页

作者： Lin Jiang Yueheng Peng Runyang He Qingqing Yang Chanlin Yi Yuqin Li Bin Zhu Yajing Si Tao Zhang Bharat B.Biswal Dezhong Yao Lan Xiong Fali Li Peng Xu The Clinical Hospital of Chengdu Brain Science Institute MOE Key Lab for NeuroinformationUniversity of Electronic Science and Technology of ChinaChengdu 611731China School of Life Science and Technology Center for Information in BioMedicineUniversity of Electronic Science and Technology of ChinaChengdu 611731China School of Psychology Xinxiang Medical UniversityXinxiang 453003China School of Science Xihua UniversityChengdu 610039China Department of Biomedical Engineering New Jersey Institute of TechnologyNewarkNJUSA School of Electrical Engineering Zhengzhou UniversityZhengzhou 450001China Research Unit of NeuroInformation Chinese Academy of Medical Sciences2019RU035ChengduChina Montreal Neurological Institute and Hospital McGill UniversityMontrealQCCanada Department of Electrical and Computer Engineering Faculty of Science and TechnologyUniversity of MacaoMacaoChina Radiation Oncology Key Laboratory of Sichuan Province 610041 ChengduChina Rehabilitation Center Qilu Hospital of Shandong UniversityJinan 250012China

Human cognition is usually underpinned by intrinsic structure and functional neural co-activation in spatially distributed brain *** to lacking an effective approach to quantifying the covarying of structure and functional responses,how the structural–functional circuits interact and how genes encode the relationships,to deepen our knowledge of human cognition and disease,are still ***,we propose a multimodal covariance network(MCN)construction approach to capture interregional covarying of the structural skeleton and transient functional activities for a single *** further explored the potential association between brain-wide gene expression patterns and structural–functional covarying in individuals involved in a gambling task and individuals with major depression disorder(MDD),adopting multimodal data from a publicly available human brain transcriptomic atlas and 2 independent *** analysis showed a replicable cortical structural–functional fine map in healthy individuals,and the expression of cognition-and disease phenotype-related genes was found to be spatially correlated with the corresponding MCN *** analysis of cell type-specific signature genes suggests that the excitatory and inhibitory neuron transcriptomic changes could account for most of the observed correlation with task-evoked MCN differences.

关键词： alterations evoked spatially

来源：评论

学校读者我要写书评

暂无评论

Collaboration-Aware Hybrid Learning for Knowledge Development Prediction 24

Collaboration-Aware Hybrid Learning for Knowledge Developmen...

引用

33rd ACM Web Conference, WWW 2024

作者： Chen, Liyi Qin, Chuan Sun, Ying Song, Xin Xu, Tong Zhu, Hengshu Xiong, Hui School of Data Science University of Science and Technology of China Hefei China Career Science Lab BOSS Zhipin Beijing China PBC School of Finance Tsinghua University Beijing China Guangzhou China Baidu Inc. Beijing China Department of Computer Science and Engineering The Hong Kong University of Science and Technology Hong Kong

ISBN: (纸本)9798400701719

In recent years, the rise of online Knowledge Management Systems (KMSs) has significantly improved work efficiency in enterprises. Knowledge development prediction, as a critical application within these online platforms, enables organizations to proactively address knowledge gaps and align their learning initiatives with evolving job requirements. However, it still confronts challenges in exploring the influence of collaborative networks on knowledge development and adapting to ecological situations in working environment. To this end, in this paper, we propose a Collaboration-Aware Hybrid Learning approach (CAHL) for predicting the future knowledge acquisition of employees and quantifying the impact of various knowledge learning patterns. Specifically, to fully harness the inherent rules of knowledge development, we first learn the knowledge co-occurrence and prerequisite relationships with an association prompt attention mechanism to generate effective knowledge representations through a specially-designed Job Knowledge Embedding module. Then, we aggregate the features of mastering knowledge and work collaborators for employee representations in another Employee Embedding module. Moreover, we propose to model the process of employee knowledge development via a Hybrid Learning Simulation module that integrates both collaborative learning and self learning to predict future-acquired job knowledge of employees. Finally, extensive experiments conducted on a real-world dataset clearly validate the effectiveness of CAHL. © 2024 ACM.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

DICE: Dynamically Induced Cross Entropy for Robust Learning with Noisy labels 6th

DICE: Dynamically Induced Cross Entropy for Robust Learning ...

引用

6th International Conference on Cognitive Systems and Signal Processing, ICCSIP 2021

作者： Liu, Tianyu State Key Lab on Intelligence Technology and Systems Department of Computer Science and Technology Tsinghua University Beijing China

ISBN: (纸本)9789811692468

Image classification has lead to the revolution of artificial intelligence in the past five years. However, image classification algorithms are significantly affected by the inherent variance in sensory input and noise in the labelled data in a real-world situation. Also, the class count of items in the real world is significantly larger than that of the typical experiment setup. How to speedily train/fine-tune a large scale model in giant label space with considerable noise is a recent interest of the machine learning community. This paper proposed a multi-stage training algorithm to fine-tune a pre-trained EfficientNet model on AliProducts large scale product classification dataset, which has a large label space (50030 classes) and severe label noise. Our method can generalize well on such a dataset while keeping the prior knowledge gained in the large-scale pretraining stage. With our novel Dynamically Induced Cross-Entropy(DICE) network loss and several other methods to tackle unbalanced datasets and improve model convergence, the model achieved 76.67% of Top-1 Accuracy and 85.42% of Top-5 accuracy, which are 3.63%/2.9% higher than symmetric cross-entropy and are significantly higher than the usual fine-tuning method with categorical cross-entropy loss(CCE). © 2022, Springer Nature Singapore Pte Ltd.

关键词： Image classification

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：