检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Kim, Eunji Mantena, Sriya Yang, Weiwei Singh, Chandan Yoon, Sungroh Gao, Jianfeng Microsoft Research United States Department of Electrical and Computer Engineering Seoul National University Korea Republic of Stanford University United States Interdisciplinary Program in Artificial Intelligence Seoul National University Korea Republic of

Recent large language models (LLMs) have excelled across a wide range of tasks, but their use in high-stakes and compute-limited settings has intensified the demand for interpretability and efficiency. We address this need by proposing Induction-head ngram models (Induction-Gram), a method that builds an efficient, interpretable LM by bolstering modern ngram models with a hand-engineered "induction head". This induction head uses a custom neural similarity metric to efficiently search the model’s input context for potential next-word completions. This process enables Induction-Gram to provide ngram-level grounding for each generated token. Moreover, experiments show that this simple method significantly improves next-word prediction over baseline interpretable models (up to 26%p) and can be used to speed up LLM inference for large models through speculative decoding. We further study Induction-Gram in a natural-language neuroscience setting, where the goal is to predict the next fMRI response in a sequence. It again provides a significant improvement over interpretable models (20% relative increase in the correlation of predicted fMRI responses), potentially enabling deeper scientific investigation of language selectivity in the brain. The code is available at https://***/ejkim47/induction-gram. © 2024, CC BY-NC-SA.

关键词： Digital elevation model

来源：评论

学校读者我要写书评

暂无评论

Investigating the Impact of Data Augmentation for Fine-Grained Grocery Product Classification Based on Vision Transformer

Investigating the Impact of Data Augmentation for Fine-Grain...

引用

Artificial Intelligence Robotics, Signal and Image Processing (AIRoSIP), International Conference on

作者： Rissa Rahmania Yaya Heryadi Lukas Lukas Wayan Suparta Ilvico Sonata Computer Science Department BINUS Graduate Program Computer Science Bina Nusantara University Jakarta Indonesia Cognitive Engineering Group (CEG) Faculty of Engineering Universitas Katolik Atma Jaya Jakarta Indonesia Department of Electrical Engineering Faculty of Industrial Technology Institut Teknologi Nasional Yogyakarta Yogyakarta Indonesia

ISBN: (数字)9798350327472

ISBN: (纸本)9798350327489

The existence of fine-grained image classification supporting smart retail provides effectiveness in recognizing products with high similarity. However, the generic classification method performs poorly in identifying products from a subordinate category. This paper aims to identify augmentation techniques to leverage the Vision Transformer (ViT) model to classify the fine-grained grocery product, which involves embedding patches and transformer encoders to extract the main features. First, we develop a fine-grained image dataset with ColorJitter, CutOut, and combining both augmentations. Secondly, we perform experiments and analysis of ViT size, embedded patch size and image size in the patch embedding process. Lastly, the ViT model are evaluated according to the image sizes 224, 384, and 512 in accuracy, loss, and confusion matrix. The highest accuracy was obtained at 0.9922. The ColorJitter and CutOut improved the confusion matrix in ViT-B/16 and ViT-L/16 with an image size of 384 and 512. The results show that both augmentations in the ViT model are able to distinguish fine-grained grocery products.

关键词： Surveys computer vision Accuracy Image recognition Transformers Feature extraction Data augmentation Artificial intelligence Image classification

来源：评论

学校读者我要写书评

暂无评论

The Effect of Major Depressive Disorder on Producing Intelligible Speech

SSRN

引用

SSRN 2024年

作者： Seneviratne, Nadee Li, Leslie Espy-Wilson, Carol Department of Electrical and Computer Engineering University of Maryland College ParkMD20740 United States Program in Neuroscience and Cognitive Science Department of Linguistics University of Maryland College ParkMD20740 United States

Understanding the characteristics of depressed speech can provide insights on aspects to consider when developing biomarkers for automatic depression detection systems. We present the results of a speech intelligibility study that was conducted to assess the intelligibility of the speech of a female and a male subject when they were severely depressed and when they were in remission. Our analysis shows that, in general, the depressed speech is at least as intelligible as in-remission speech. This is further supported by the underlying "simpler" articulatory coordination of depressed speech as derived from the Articulatory Coordination Features. Acoustically, the simpler articulatory coordination means that there is less variability due to coarticulation in the speech signal. © 2024, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Thermal energy dependent transient permittivity of epsilon-near-zero material

引用

Science China(Physics,Mechanics & Astronomy) 2022年第8期65卷 96-101页

作者： Heng Wang Lixun Sun Kang Du Wending Zhang SooJin Chua Guixin Li Ting Mei MOE Key Laboratory of Material Physics and Chemistry under Extraordinary Conditions and Shaanxi Key Laboratory of Optical Information TechnologySchool of Physical Science and TechnologyNorthwestern Polytechnical UniversityXi’an710129China Department of Materials Science and Engineering Southern University of Science and TechnologyShenzhen518055China Shenzhen Institute for Quantum Science and Engineering Southern University of Science and TechnologyShenzhen518055China Department of Electrical and Computer Engineering National University of SingaporeSingapore117583Singapore LEES Program Singapore-MIT Alliance for Research&Technology(SMART)Singapore138602Singapore

Transparent conductive oxides exhibit attractive optical nonlinearity with ultrafast response and giant refractive index change near the epsilon-near-zero(ENZ) wavelength, originating from the intraband dynamics of conduction electrons. The optical nonlinearity of ENZ materials has been explained by using the overall-effective-mass and the overall-scattering-time of electrons in the extended Drude model. However, their response to optical excitation is yet the last building block to complete the theory. In this paper, the concept of thermal energy is theoretically proposed to account for the total energy of conduction electrons exceeding their thermal equilibrium value. The time-varying thermal energy is adopted to describe the transient optical response of indium-tin-oxide(ITO), a typical ENZ material. A spectrally-resolved femtosecond pump-probe experiment was conducted to verify our theory. By correlating the thermal energy with the pumping density, both the giant change and the transient response of the permittivity of ITO can be predicted. The results in this work provide a new methodology to describe the transient permittivities of ENZ materials, which will benefit the design of ENZ-based nonlinear photonic devices.

关键词： epsilon-near-zero optical nonlinearity transient response thermal energy hot electron indium tin oxide

来源：评论

学校读者我要写书评

暂无评论

Tone Value Representation for computer-Assisted Pronunciation Training

Tone Value Representation for Computer-Assisted Pronunciatio...

引用

12th International Conference on Speech Prosody, Speech Prosody 2024

作者： Li, Wu-Hao Liu, Te-Hsin Chiang, Chen-Yu College of Electrical Engineering and Computer Science National Taipei University Taiwan Graduate Program of Teaching Chinese as the Second Language National Taiwan University Taiwan Dept. of Communication Engineering National Taipei University New Taipei City Taiwan

Scholars have long demonstrated the effectiveness of using visual representations to aid non-native speakers in learning Mandarin tone. Among these methods, presenting pitch contours visually has proven to be the most intuitive and efficient. However, the variability of pitch contours in continuous speech, influenced by coarticulation, poses challenges for learners to follow accurately. To address this issue, we propose a method to normalize pitch contours, mitigating the effects of coarticulation and prosodic phrasing. Experimental results show that subjects can identify the class of tone by looking at the representation proposed in this study and evaluating the quality of the tones of the syllables pronounced by the speakers visually. In addition, our approach can present the mispronunciation with the contours of the relative pitch of the syllables. We believe that our method offers more comprehensive feedback to learners compared to relying solely on a tone recognition model for pronunciation evaluation. © 2024 International Speech Communications Association. All rights reserved.

关键词： Continuous speech recognition

来源：评论

学校读者我要写书评

暂无评论

MathReader: Text-to-Speech for Mathematical Documents

arXiv

引用

arXiv 2025年

作者： Hyeon, Sieun Jung, Kyudan Kim, Nam-Joon Ryu, Hyun Gon Do, Jaeyoung Department of Electrical and Computer Engineering Seoul National University Korea Republic of Department of Mathematics Chung-Ang University Korea Republic of NVIDIA United States Interdisciplinary Program in Artificial Intelligence Seoul National University Korea Republic of

TTS (Text-to-Speech) document reader from Microsoft, Adobe, Apple, and OpenAI have been serviced worldwide. They provide relatively good TTS results for general plain text, but sometimes skip contents or provide unsatisfactory results for mathematical expressions. This is because most modern academic papers are written in LaTeX, and when LaTeX formulas are compiled, they are rendered as distinctive text forms within the document. However, traditional TTS document readers output only the text as it is recognized, without considering the mathematical meaning of the formulas. To address this issue, we propose MathReader, which effectively integrates OCR, a fine-tuned T5 model, and TTS. MathReader demonstrated a lower Word Error Rate (WER) than existing TTS document readers, such as Microsoft Edge and Adobe Acrobat, when processing documents containing mathematical formulas. MathReader reduced the WER from 0.510 to 0.281 compared to Microsoft Edge, and from 0.617 to 0.281 compared to Adobe Acrobat. This will significantly contribute to alleviating the inconvenience faced by users who want to listen to documents, especially those who are visually impaired. The code is available at https://***/hyeonsieun/MathReader. Copyright © 2025, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

ReViT: A Hybrid Approach for BCLC Staging of Hepatocellular Carcinoma Using 3D CT with Multiple Instance Learning

ReViT: A Hybrid Approach for BCLC Staging of Hepatocellular ...

引用

2024 IEEE-EMBS International Conference on Biomedical and Health Informatics, BHI 2024

作者： Chang, Shun-Cheng Yu, Hsin-Pei Hsieh, Yi-Hsien Wang, Pochuang Wang, Weichung Su, Tung-Hung Kao, Jia-Horng Lin, Che Taipei10617 Taiwan Department of Computer Science and Information Engineering Ntu Taipei10617 Taiwan Institute of Applied Mathematical Sciences Ntu Taipei10617 Taiwan National Taiwan University Hospital Division of Gastroenterology and Hepatology Department of Internal Medicine Taipei10048 Taiwan Department of Electrical Engineering Ntu Taipei10617 Taiwan Center for Advanced Computing and Imaging in Biomedicine Ntu Taipei10617 Taiwan Smart Medicine and Health Informatics Program Ntu Taipei10617 Taiwan

ISBN: (纸本)9798350351552

Deep learning has revolutionized medical imaging, offering advanced methods for accurate diagnosis and treatment planning. The BCLC staging system is crucial for staging Hepatocellular Carcinoma (HCC), a high-mortality cancer. An automated BCLC staging system could significantly enhance diagnosis and treatment planning efficiency. However, we found that BCLC staging, which is directly related to the size and number of liver tumors, aligns well with the principles of the Multiple Instance Learning (MIL) framework. To effectively achieve this, we proposed a new preprocessing technique called Masked Cropping and Padding(MCP), which addresses the variability in liver volumes and ensures consistent input sizes. This technique preserves the structural integrity of the liver, facilitating more effective learning. Furthermore, we introduced Re ViT, a novel hybrid model that integrates the local feature extraction capabilities of Convolutional Neural Networks (CNNs) with the global context modeling of Vision Transformers (ViTs). Re ViT leverages the strengths of both architectures within the MIL framework, enabling a robust and accurate approach for BCLC staging. We will further explore the trade-off between performance and interpretability by employing TopK Pooling strategies, as our model focuses on the most informative instances within each bag. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

A BCLC Staging System for Hepatocellular Carcinoma using Swin Transformer and CT Imaging 46

A BCLC Staging System for Hepatocellular Carcinoma using Swi...

引用

46th Annual International Conference of the IEEE engineering in Medicine and Biology Society, EMBC 2024

作者： Chang, Shun-Cheng Wang, Pochuang Wang, Weichung Su, Tung-Hung Kao, Jia-Horng Lin, Che Taipei10617 Taiwan Department of Computer Science and Information Engineering Ntu Taipei10617 Taiwan Institute of Applied Mathematical Sciences Ntu Taipei10617 Taiwan Division of Gastroenterology and Hepatology Department of Internal Medicine National Taiwan University Hospital Taipei10048 Taiwan Department of Electrical Engineering Ntu Taipei10617 Taiwan Center for Advanced Computing and Imaging in Biomedicine Ntu Taipei10617 Taiwan Smart Medicine and Health Informatics Program Ntu Taipei10617 Taiwan

ISBN: (纸本)9798350371499

The Barcelona Clinic Liver Cancer (BCLC) staging system plays a crucial role in clinical planning, offering valuable insights for effectively managing hepatocellular carcinoma. Accurate prediction of BCLC stages can significantly ease the workload on radiologists. However, few datasets are explicitly designed for discerning BCLC stages. Despite the common practice of appending BCLC labels to clinical data within datasets, the inherent imbalance in BCLC distribution is further amplified by the diverse purposes for which datasets are curated. In this study, we aim to develop a BCLC staging system using the advanced Swin Transformer model. Additionally, we explore the integration of two datasets, each originally intended for separate objectives, highlighting the critical challenge of preserving class distribution in practical study designs. This exploration is pivotal for ensuring the applicability of our developed staging system in the designed clinical settings. Our resulting BCLC staging system demonstrates an accuracy of 55.81% (±7.8%), contributing to advancing medical image-based research for predicting BCLC stages. © 2024 IEEE.

关键词： Distribution transformers

来源：评论

学校读者我要写书评

暂无评论

Fuzzy Logic Approach for Power Transformer Condition Assessment Using Health Index

Fuzzy Logic Approach for Power Transformer Condition Assessm...

引用

2022 International Conference on Power, Energy and Innovations, ICPEI 2022

作者： Dorji, Tashi Suwanasri, Thanapong Poonnoy, Nitchamon Suwanasri, Cattareeya EPE Program The Sirindhorn International Thai-German Graduate School of Engineering KMUTNB 1518 Pracharat 1 Rd. Bangsue Bangkok Thailand Department of Teacher Training in Electrical Engineering Faculty of Technical Education KMUTNB 1518 Pracharat 1 Rd. Bangsue Bangkok Thailand Department of Electrical and Computer Engineering Faculty of Engineering KMUTNB 1518 Pracharat 1 Rd.Bangsue Bangkok Thailand

ISBN: (数字)9781665460194

ISBN: (纸本)9781665460194

This paper presents a realistic power transformer condition assessment and Health Index (HI) determination technique. The real-time operating data of twenty-seven power transformers of Bhutan Power Corporation Limited in transmission substations are used for condition evaluation. The diagnostic test data are combined into nine diagnostic groups for condition assessment. The conventional method of scoring and weighting average and the fuzzy logic methods are adopted for the condition assessment of power transformers. The excel program.was developed for the conventional method for seamless evaluation. All the parameters harmonies the estimation and classification of the transformer operating condition. Then, this paper proposed new procedures for validating the conventional system of HI determination by the fuzzy logic method. © 2022 IEEE.

关键词： Power transformers

来源：评论

学校读者我要写书评

暂无评论

Enabling AI-Generated Content (AIGC) Services in Wireless Edge Networks

arXiv

引用

arXiv 2023年

作者： Du, Hongyang Li, Zonghang Niyato, Dusit Kang, Jiawen Xiong, Zehui Shen, Xuemin Kim, Dong In The School of Computer Science and Engineering The Energy Research Institute @ NTU Interdisciplinary Graduate Program Nanyang Technological University Singapore The School of Information and Communication Engineering University of Electronic Sciences and Technology of China Chengdu China The School of Automation Guangdong University of Technology China The Pillar of Information Systems Technology and Design Singapore University of Technology and Design Singapore The Department of Electrical and Computer Engineering University of Waterloo Canada The Department of Electrical and Computer Engineering Sungkyunkwan University Korea Republic of

Artificial Intelligence-Generated Content (AIGC) refers to the use of AI to automate the information creation process while fulfilling the personalized requirements of users. However, due to the instability of AIGC models, e.g., the stochastic nature of diffusion models, the quality and accuracy of the generated content can vary significantly. In wireless edge networks, the transmission of incorrectly generated content may unnecessarily consume network resources. Thus, a dynamic AIGC service provider (ASP) selection scheme is required to enable users to connect to the most suited ASP, improving the users’ satisfaction and quality of generated content. In this article, we first review the AIGC techniques and their applications in wireless networks. We then present the AIGC-as-a-service (AaaS) concept and discuss the challenges in deploying AaaS at the edge networks. Yet, it is essential to have performance metrics to evaluate the accuracy of AIGC services. Thus, we introduce several image-based perceived quality evaluation metrics. Then, we propose a general and effective model to illustrate the relationship between computational resources and user-perceived quality evaluation metrics. To achieve efficient AaaS and maximize the quality of generated content in wireless edge networks, we propose a deep reinforcement learning-enabled algorithm for the optimal ASP selection. Simulation results show that the proposed algorithm can provide a higher quality of generated content to users and achieve fewer crashed tasks by comparing with four benchmarks, i.e., overloading-avoidance, random, round-robin policies, and the upper-bound schemes. © 2023, CC BY.

关键词： Reinforcement learning

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：