检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Li, Pengxiang Gao, Zhi Zhang, Bofei Yuan, Tao Wu, Yuwei Harandi, Mehrtash Jia, Yunde Zhu, Song-Chun Li, Qing Beijing Key Laboratory of Intelligent Information Technology School of Computer Science & Technology Beijing Institute of Technology China State Key Laboratory of General Artificial Intelligence BIGAI China State Key Laboratory of General Artificial Intelligence Peking University China Department of Electrical and Computer System Engineering Monash University Australia Guangdong Laboratory of Machine Perception and Intelligent Computing Shenzhen MSU-BIT University China Department of Automation Tsinghua University China

Vision language models (VLMs) have achieved impressive progress in diverse applications, becoming a prevalent research direction. In this paper, we build FIRE, a feedback-refinement dataset, consisting of 1.1M multi-turn conversations that are derived from 27 source datasets, empowering VLMs to spontaneously refine their responses based on user feedback across diverse tasks. To scale up the data collection, FIRE is collected in two components: FIRE-100K and FIRE-1M, where FIRE-100K is generated by GPT-4V, and FIRE-1M is freely generated via models trained on FIRE-100K. Then, we build FIRE-Bench, a benchmark to comprehensively evaluate the feedback-refining capability of VLMs, which contains 11K feedback-refinement conversations as the test data, two evaluation settings, and a model to provide feedback for VLMs. We develop the FIRE-LLaVA model by fine-tuning LLaVA on FIRE-100K and FIRE-1M, which shows remarkable feedback-refining capability on FIRE-Bench and outperforms untrained VLMs by 50%, making more efficient user-agent interactions and underscoring the significance of the FIRE dataset. Copyright © 2024, The Authors. All rights reserved.

关键词： Fireclay

来源：评论

学校读者我要写书评

暂无评论

Molecular dynamics study on bubble nucleation characteristics on rectangular Nano-Grooved surfaces

引用

Journal of Molecular Liquids 2025年 420卷

作者： Li, Zhibin Lou, Jiacheng Wu, Xiaoyi Li, Xujun Chang, Fucheng Wang, Hengyuan Li, Huixiong State Key Laboratory of Multiphase Flow in Power Engineering Xi'an Jiaotong University Xi'an710049 China National Key Lab of Aerospace Power System and Plasma Technology School of Mechanical Engineering Xi'an Jiaotong University Xi'an710049 China

The micro/nanoscale structures on metal surfaces are of significant research interest for boiling heat transfer and exhibit great application potential in the thermal management of high-power electronic devices. This study employs the molecular dynamics simulation method to investigate the boiling heat transfer characteristics of hydrophilic rectangular shallow nano-groove structures, comparing them with the smooth surface. The phase transition heat transfer characteristics of liquid films on the smooth surface, single rectangular groove (SRG) surface, and overlapping rectangular groove (ORG) surface are revealed by analyzing the variations of parameters such as energy contour diagrams, bubble volume, liquid film temperature, and average potential energy. The research results indicate that water molecules are in a solid-like state within shallow grooves, while the nucleation sites of initial bubbles form randomly outside the grooves. During the initial heating and bubble growth, the initial nucleation time and bubble growth rate of ORG are slower than those of SRG, due to the small rectangular groove on ORG that diminishes the energy transfer efficiency. During bubble formation and growth, the average kinetic energy within the liquid film fluctuates at the nucleation moment in the critical nucleation region, and the average potential energy rises from −0.083 eV. These findings provide a theoretical basis for the micro/nanoscale structure design of vapor-liquid phase transition thermal management devices. © 2024 Elsevier B.V.

关键词： Kinetic energy

来源：评论

学校读者我要写书评

暂无评论

PROJECTION HEAD IS SECRETLY AN INFORMATION BOTTLENECK

arXiv

引用

arXiv 2025年

作者： Ouyang, Zhuo Hu, Kaiwen Zhang, Qi Wang, Yifei Wang, Yisen College of Engineering Peking University China School of EECS Peking University China State Key Lab of General Artificial Intelligence School of Intelligence Science and Technology Peking University China MIT CSAIL United States Institute for Artificial Intelligence Peking University China

Recently, contrastive learning has risen to be a promising paradigm for extracting meaningful data representations. Among various special designs, adding a projection head on top of the encoder during training and removing it for downstream tasks has proven to significantly enhance the performance of contrastive learning. However, despite its empirical success, the underlying mechanism of the projection head remains under-explored. In this paper, we develop an in-depth theoretical understanding of the projection head from the information-theoretic perspective. By establishing the theoretical guarantees on the downstream performance of the features before the projector, we reveal that an effective projector should act as an information bottleneck, filtering out the information irrelevant to the contrastive objective. Based on theoretical insights, we introduce modifications to projectors with training and structural regularizations. Empirically, our methods exhibit consistent improvement in the downstream performance across various real-world datasets, including CIFAR-10, CIFAR-100, and ImageNet-100. We believe our theoretical understanding on the role of the projection head will inspire more principled and advanced designs in this field. Code is available at https://***/PKU-ML/Projector_Theory. © 2025, CC BY.

关键词： Information theory

来源：评论

学校读者我要写书评

暂无评论

An optimized multi-modular resonant switched capacitor converter 4

An optimized multi-modular resonant switched capacitor conve...

引用

2022 4th International Conference on Energy, Power and Grid, ICEPG 2022

作者： Li, Ruishu Chang, Xiao Han, Xiaoqing He, Xin Guo, Zhuoyan Meng, Runquan Shanxi Key Lab of Power System Operation and Control Taiyuan University of Technology Taiyuan China Electric Power Research Institute State Grid Shanxi Electric Power Company Taiyuan China

The switched capacitor converters have the advantages of high efficiency and high power density, which are widely used in data centers, distributed photovoltaic power generation, and other fields. In this paper, an optimized multi-modular resonant switched capacitor converter (MMRSCC) is proposed for an intermediate bus converter in the power supply system of data centers. By removing the inductance of part of the coupled resonant circuit, the converter can eliminate the influence of the parameter difference of passive components, and no additional voltage clamp circuit is needed, which improves the power density. At the same time, the half-bridge with the same switching action on the rectifier side is combined into one, which reduces the number of switching devices and is more conducive to improving the voltage conversion efficiency. In addition, the inductor of the basic unit of the resonant switched capacitor with the same state is shared to reduce the volume of the passive device, further improve the power density, and optimize the MMRSCC. Finally, the correctness of the topology is verified by simulation on the PLECS. © Published under licence by IOP Publishing Ltd.

关键词： Conversion efficiency

来源：评论

学校读者我要写书评

暂无评论

Unsupervised Semantic Segmentation with Contrastive Translation Coding 6th

Unsupervised Semantic Segmentation with Contrastive Translat...

引用

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

作者： Chen, Runfa Sun, Hanbing Wang, Ling State Key Lab on Intelligent Technology and Systems Department of Computer Science and Technology Tsinghua University Beijing China Department of Air Defense Early Warning Command System Air Force Early Warning Academy Wuhang China

ISBN: (纸本)9789811692468

The proposed new method constructs a contrastive learning task through unsupervised image-to-image translation and indirectly extracts domain-invariant features by maximizing mutual information. Specifically, by reusing the semantic segmentation network for image-to-image translation, the probability space of the segmentation output is projected into the RGB space, thereby constructing positive and negative samples to maximize the mutual information between the input image and the semantic segmentation RGB projection map. Meanwhile, the output space of the target domain is transferred to the source domain, and then the robust domain-invariant semantic features are extracted. We develop two adversarial transfer methods and a three-stage training paradigm of pre-training, cross-domain transfer, and self-supervised training. Experiments on benchmarks demonstrate that this method is reasonable and feasible. Comprehensive ablation studies and analyses are also carried out to reveal the advantages and disadvantages of the two designed transfer schemes and the effect and significance of color-invariant semantic information for unsupervised semantic segmentation tasks. The contrastive learning task, which is constructed through unsupervised cross-domain image-to-image translation, provides a new insight for cross-domain transfer learning to overcome the problem of domain shift. © 2022, Springer Nature Singapore Pte Ltd.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Series-Connected SiC MOSFETs Module with a Single Gate Driver 6

Series-Connected SiC MOSFETs Module with a Single Gate Drive...

引用

6th IEEE Conference on Energy Internet and Energy system Integration, EI2 2022

作者： Zhang, Yuxuan Li, Rui Meng, Runquan Ren, Yuhan Ma, Yue Chen, Juan Taiyuan University of Technology Shanxi Key Lab of Power System Operation and Control Taiyuan China Electric Power Research Institute of State Grid Shanxi Electric Power Company Taiyuan China

ISBN: (纸本)9798350347159

With the continuous development and progress of power electronics technology, the technological innovation of each generation of power semiconductors will have a profound impact on the generation, transmission, conversion and distribution of electrical energy, and has made outstanding contributions to the advancement of human civilization and economic and social development. In recent years, a new generation of wide-bandgap semiconductor power devices has developed vigorously, and new material switching devices such as SiC and GaN MOSFET have entered people's field of vision, which has brought new historical opportunities for the performance improvement of power electronic equipment, as well as new challenges. This paper presents a series-connected SiC MOSFET switching module topology with a single gate driver. Then, the working principle including static state and switching transition is analyzed. The LTspice simulations in the third section validate the analysis and show the usability of the series-connected SiC MOSFET stack topology using a single gate driver. © 2022 IEEE.

关键词： Topology

来源：评论

学校读者我要写书评

暂无评论

De-confounded variational encoder-decoder for logical table-to-text generation 59

De-confounded variational encoder-decoder for logical table-...

引用

Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, ACL-IJCNLP 2021

作者： Chen, Wenqing Tian, Jidong Li, Yitian He, Hao Jin, Yaohui MoE Key Lab of Artificial Intelligence AI Institute Shanghai Jiao Tong University China State Key Lab of Advanced Optical Communication System and Network Shanghai Jiao Tong University China

ISBN: (纸本)9781954085527

Logical table-to-text generation aims to automatically generate fluent and logically faithful text from tables. The task remains challenging where deep learning models often generated linguistically fluent but logically inconsistent text. The underlying reason may be that deep learning models often capture surface-level spurious correlations rather than the causal relationships between the table x and the sentence y. Specifically, in the training stage, a model can get a low empirical loss without understanding x and use spurious statistical cues instead. In this paper, we propose a de-confounded variational encoder-decoder (DCVED) based on causal intervention, learning the objective p(y|do(x)). Firstly, we propose to use variational inference to estimate the confounders in the latent space and cooperate with the causal intervention based on Pearl's do-calculus to alleviate the spurious correlations. Secondly, to make the latent confounder meaningful, we propose a back-prediction process to predict the not-used entities but linguistically similar to the exactly selected ones. Finally, since our variational model can generate multiple candidates, we train a table-text selector to find out the best candidate sentence for the given table. An extensive set of experiments show that our model outperforms the baselines and achieves new state-of-the-art performance on two logical table-to-text datasets in terms of logical fidelity. © 2021 Association for Computational Linguistics

关键词： Calculations

来源：评论

学校读者我要写书评

暂无评论

Convergence Rate Analysis of LION

arXiv

引用

arXiv 2024年

作者： Dong, Yiming Li, Huan Lin, Zhouchen State Key Lab of General Artificial Intelligence School of Intelligence Science and Technology Peking University Beijing100871 China Institute of Robotics and Automatic Information Systems College of Artificial Intelligence Nankai University Tianjin300071 China Guangzhou510555 China

The LION (evoLved sIgn mOmeNtum) optimizer for deep neural network training was found by Google via program search, with the simple sign update yet showing impressive performance in training large scale networks. Although previous studies have investigated its convergence properties, a comprehensive analysis, especially the convergence rate, is still desirable. Recognizing that LION can be regarded as solving a specific constrained problem, this paper focuses on demonstrating its convergence to the Karush-Kuhn-Tucker (KKT) point at the rate of O(√dK-1/4) measured by gradient 1 norm, where d is the problem dimension and K is the number of iteration steps. Step further, we remove the constraint and establish that LION converges to the critical point of the general unconstrained problem at the same rate. This rate not only delivers the currently optimal dependence on the problem dimension d but also tightly matches the theoretical lower bound for nonconvex stochastic optimization algorithms, which is typically measured using the gradient 2 norm, with respect to the number of iterations K. Through extensive experiments, we not only demonstrate that LION achieves lower loss and higher performance compared to standard SGD, but also empirically confirm that the gradient 1/2 norm ratio aligns with Θ(√d), thus proving that our convergence rate matches the theoretical lower bound with respect to d in the empirical sense. Copyright © 2024, The Authors. All rights reserved.

关键词： Stochastic systems

来源：评论

学校读者我要写书评

暂无评论

Progressive Residual Extraction Based Pre-Training for Speech Representation Learning

IEEE Transactions on Audio, Speech and Language Processing

引用

IEEE Transactions on Audio, Speech and Language Processing 2025年 33卷 1825-1837页

作者： Tianrui Wang Jin Li Ziyang Ma Rui Cao Xie Chen Longbiao Wang Meng Ge Xiaobao Wang Yuguang Wang Jianwu Dang Nyima Tashi Tianjin Key Laboratory of Cognitive Computing and Application College of Intelligence and Computing Tianjin University Tianjin China MoE Key Lab of Artificial Intelligence AI Institute Shanghai Jiao Tong University Shanghai China Huiyan Technology (Tianjin) Company Ltd. Tianjin China Saw Swee Hock School of Public Health National University of Singapore Singapore Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen China School of Information Science and Technology Tibet University Lhasa China State Key Laboratory of Tibetan Intelligence Lhasa China

Self-supervised learning (SSL) has garnered significant attention in speech processing, particularly excelling in linguistic tasks such as speech recognition. However, improving the performance of pre-trained models across various downstream tasks—each requiring distinct types of speech information—remains a significant challenge. To address this, we propose a progressive residual extraction based SSL method, named ProgRE. Specifically, we introduce two lightweight, specialized task modules into an encoder-style SSL backbone to enhance its ability to extract pitch variation and speaker information from speech. Furthermore, to mitigate the incompatibility between the reinforced pitch variation and speaker information and the learning of content information, we employ residual extraction, leveraging the extracted representations as references or conditioning signals to guide the subsequent modules in more effectively learning content-related information under the supervision of HuBERT-based speech masking prediction. In this manner, we can incrementally extract pitch variation, speaker, and content representations from the input speech. Finally, these multiple representations, each capturing diverse speech information, are combined using different layer weights to produce task-specific representations for various downstream tasks. Experimental results demonstrate that our ProgRE achieves significant performance improvements across several tasks, such as speaker identification, speech recognition, emotion recognition, speech enhancement, and voice conversion, outperforming excellent SSL methods like wav2vec2.0, HuBERT, and WavLM.

关键词： Data mining Feature extraction Transformers Speech recognition Training Speech enhancement Acoustics Vector quantization Self-supervised learning Germanium

来源：评论

学校读者我要写书评

暂无评论

Improving the Sparse Structure Learning of Spiking Neural Networks from the View of Compression Efficiency

arXiv

引用

arXiv 2025年

作者： Shen, Jiangrong Xu, Qi Pan, Gang Chen, Badong Faculty of Electronic and Information Engineering Xi’an Jiaotong University China Institute of Artificial Intelligence and Robotics Xi’an Jiaotong University China State Key Lab of Brain-Machine Intelligence Zhejiang University China National Key Lab of Human-Machine Hybrid Augmented Intelligence Xi’an Jiaotong University China School of Computer Science Dalian University of Technology China

The human brain utilizes spikes for information transmission and dynamically reorganizes its network structure to boost energy efficiency and cognitive capabilities throughout its lifespan. Drawing inspiration from this spike-based computation, Spiking Neural Networks (SNNs) have been developed to construct event-driven models that emulate this efficiency. Despite these advances, deep SNNs continue to suffer from over-parameterization during training and inference, a stark contrast to the brain’s ability to self-organize. Furthermore, existing sparse SNNs are challenged by maintaining optimal pruning levels due to a static pruning ratio, resulting in either under or over-pruning. In this paper, we propose a novel two-stage dynamic structure learning approach for deep SNNs, aimed at maintaining effective sparse training from scratch while optimizing compression efficiency. The first stage evaluates the compressibility of existing sparse subnetworks within SNNs using the PQ index, which facilitates an adaptive determination of the rewiring ratio for synaptic connections based on data compression insights. In the second stage, this rewiring ratio critically informs the dynamic synaptic connection rewiring process, including both pruning and regrowth. This approach significantly improves the exploration of sparse structures training in deep SNNs, adapting sparsity dynamically from the point view of compression efficiency. Our experiments demonstrate that this sparse training approach not only aligns with the performance of current deep SNNs models but also significantly improves the efficiency of compressing sparse SNNs. Crucially, it preserves the advantages of initiating training with sparse models and offers a promising solution for implementing Edge AI on neuromorphic hardware. © 2025, CC BY-NC-ND.

关键词：

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：