检索结果-内蒙古大学图书馆

Joint Workshop of the 9th Financial technology and Natural Language Processing, FinNLP 2025, the 6th Financial Narrative Processing, FNP 2025, and the 1st Workshop on Large Language Models for Finance and Legal, LLMFinLegal 2025, co-located with the 31st International Conference on Computational Linguistics, COLING 2025

作者： Nishida, Shunsuke Utsuro, Takehito Degree Programs in Systems and Information Engineering Graduate School of Science and Technology University of Tsukuba Japan

ISBN: (纸本)9798891762091

In this paper, we study the task of generating financial news articles related to stock price fluctuations. Traditionally, reporters manually write these articles by identifying the causes behind significant stock price volatility. However, this process is time-consuming, limiting the number of articles produced. To address this, the study explores the use of generative AI to automatically generate such articles. The AI system, similar to human reporters, would analyze stock price volatility and determine the underlying factors contributing to these fluctuations. To support this approach, we introduces a Japanese dataset called JFinSR, which includes stock price fluctuation rankings from "Kabutan" and related financial information regarding factors of stock price rise / decline from "Nihon Keizai Shimbun (Nikkei)." Using this dataset, we implement the few-shot learning technique on large language models (LLMs) to enable automatic generation of high-quality articles from factors of stock price rise / decline that are available in Nikkei. In the evaluation, we compare zero-shot and few-shot learning approaches, where the few-shot learning achieved the higher F1 scores in terms of ROUGE-1/ROUGE-L metrics. © 2025 Association for Computational Linguistics.

关键词： Zero-shot learning

来源：评论

学校读者我要写书评

暂无评论

Embedded Topic Models Enhanced by Wikification 1

Embedded Topic Models Enhanced by Wikification

引用

1st Workshop on Advancing Natural Language Processing for Wikipedia, WikiNLP 2024

作者： Shibuya, Takashi Utsuro, Takehito Degree Programs in Systems and Information Engineering Graduate School of Science and Technology University of Tsukuba Japan

ISBN: (纸本)9798891761889

Topic modeling analyzes a collection of documents to learn meaningful patterns of words. However, previous topic models consider only the spelling of words and do not take into consideration the homography of words. In this study, we incorporate the Wikipedia knowledge into a neural topic model to make it aware of named entities. We evaluate our method on two datasets, 1) news articles of New York Times and 2) the AIDA-CoNLL dataset. Our experiments show that our method improves the performance of neural topic models in generalizability. Moreover, we analyze frequent terms in each topic and the temporal dependencies between topics to demonstrate that our entity-aware topic models can capture the time-series development of topics well. © 2024 Association for Computational Linguistics.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

A deep learning framework for automatic assessment of presence in virtual reality using multimodal behavioral cues

引用

Neural Computing and Applications 2025年第8期37卷 6283-6303页

作者： Pannattee, Peerawat Shimada, Shogo Yem, Vibol Nishiuchi, Nobuyuki Graduate School of Systems Design Tokyo Metropolitan University Tokyo191-0065 Japan Graduate School of Science and Technology Degree Program in Systems and Information Engineering University of Tsukuba Ibaraki3058573 Japan

Effective development of virtual reality (VR) applications is heavily reliant on the evaluation of user experience (UX). However, traditional methods such as questionnaires have inherent limitations which hinder their ability to capture nuanced behavioral responses and impede the agility of VR content creation: They are time-consuming, burdensome for users, and require significant human effort for interpretation. This study introduces an automated framework aimed at addressing the limitations of questionnaire-based evaluations to assess UX in VR. Our primary focus is to validate the concept of this framework through assessment of the sense of presence (SOP), a crucial psychological perception with significant impact on VR UX. Our proposed framework utilizes a deep neural network (DNN) to analyze patterns in multimodal behavioral cues, including facial expressions, head movements, and hand movements, to predict scores from the Igroup Presence Questionnaire (IPQ). Additionally, we introduce two statistical profiles: the Visual Entropy Profile (VEP), which offers insights into visual complexity by depicting scene entropy, and the Experiential Presence Profile (EPP), which is designed to capture users’ historical SOP levels to enable personalized baseline and sensitivity estimation. The proposed framework achieves a significant correlation between actual and predicted IPQ scores, with a Spearman’s rank correlation coefficient of 0.7303, showcasing the potential of DNNs in analyzing complex behavioral signals and automating SOP assessment. This study represents a pioneering effort in leveraging DNNs for the automatic assessment of SOP and paves the way for future advances in automatically assessing VR UX and unlocking new opportunities in the field. © The Author(s) 2025.

关键词： Virtual environments

来源：评论

学校读者我要写书评

暂无评论

CNN Model Compression by Merit-Based Distillation 19

CNN Model Compression by Merit-Based Distillation

引用

19th IEEE International Colloquium on Signal Processing and Its Applications, CSPA 2023

作者： Morikawa, Takumi Kameyama, Keisuke University of Tsukuba Degree Programs in Systems and Information Engineering Graduate School of Science and Technology Tsukuba Japan University of Tsukuba Faculty of Engineering Information and Systems Tsukuba Japan

ISBN: (纸本)9781665476928

Deep learning has advanced dramatically in recent years, and especially large convolutional neural networks (CNNs) have shown outstanding performance in a wide variety of tasks. However, such large-scale CNNs may not be able to be utilized in practice because they require a lot of computational resources. Therefore, there is a need to realize a CNN in a small model that has high performance similar to the large CNN. One method to solve these problems is Distillation, which uses a model of large scale and high performance as the teacher and a small model as the student model, and performs model compression by transferring the knowledge of the teacher model to the student model. However, in distillation, even when the teacher model incorrectly guesses the training input, the student model learns to mimic the teacher. This can prevent the student model from learning better. In this paper, we propose Meritbased Distillation, in which the accuracy of the teacher model's output determines the degree to which the student model mimics the teacher model's output, thereby preventing the student to learn undesired mappings. In the experiment, the effect of the Merit-based Distillation is evaluated. Consequently, the proposed method outperforms the improvements in the generalization by the student models trained by other known methods. © 2023 IEEE.

关键词： Image classification

来源：评论

学校读者我要写书评

暂无评论

LSTM-based Forecasting using Policy Stringency and Time-varying Parameters of the SIR Model for COVID-19 19

LSTM-based Forecasting using Policy Stringency and Time-vary...

引用

19th IEEE International Colloquium on Signal Processing and Its Applications, CSPA 2023

作者： Maniamfu, Pavodi Kameyama, Keisuke University of Tsukuba Degree Programs in Systems and Information Engineering Graduate School of Science and Technology Tsukuba Japan University of Tsukuba Faculty of Engineering Information and Systems Tsukuba Japan

ISBN: (纸本)9781665476928

Accurate forecasting of the number of infections is an important task that can allow health care decision makers to allocate medical resources efficiently during a pandemic. Two approaches have been combined, a stochastic model by Vega et al. for modelling infectious disease and Long Short-Term Memory using COVID-19 data and government's policies. In the proposed model, LSTM functions as a nonlinear adaptive filter to modify the outputs of the SIR model for more accurate forecasts one to four weeks in the future. Our model outperforms most models among the CDC models using the United States data. We also applied the model on the Canadian data from two provinces, Saskatchewan and Ontario where it performs with a low mean absolute percentage error. © 2023 IEEE.

关键词： COVID-19

来源：评论

学校读者我要写书评

暂无评论

Machine Learning Curriculums Generated by Classifier Ensembles 19

Machine Learning Curriculums Generated by Classifier Ensembl...

引用

19th IEEE International Colloquium on Signal Processing and Its Applications, CSPA 2023

作者： Huang, Tzu-Jui Kameyama, Keisuke University of Tsukuba Degree Programs in Systems and Information Engineering Graduate School of Science and Technology Tsukuba Japan University of Tsukuba Faculty of Engineering Information and Systems Tsukuba Japan

ISBN: (纸本)9781665476928

Image classification is a fundamental task that attempts to classify the images into the classes they belong to. CNN is commonly used for the classification due to its flexibility and stability. This paper treats Curriculum Learning as an assistant method that decides the procedure the CNN ought to be trained. Curriculum Learning is generally defined as a method that divides and orders the training datasets into several different levels of learning difficulties. This paper utilizes the idea of ensemble classification that trains according to the order based on the consensus of different machine learning models. As a result, the proposed Ensemble-Generated Curriculum Learning model achieves higher accuracy compared to that of a single-stage CNN learning with dog face recognition and CIFAR-10 datasets. © 2023 IEEE.

关键词： Curricula

来源：评论

学校读者我要写书评

暂无评论

A Note on Even-Sized Spectral Clustering 13

A Note on Even-Sized Spectral Clustering

引用

Joint 13th International Conference on Soft Computing and Intelligent systems and 25th International Symposium on Advanced Intelligent systems, SCIS and ISIS 2024

作者： Miyakawa, Hiroki Endo, Yasunori Graduate School of Science and Technology Degree Programs in Systems and Information Engineering Master's Program in Risk and Resilience Engineering University of Tsukuba Ibaraki Tsukuba Japan Institute of Systems and Information Engineering University of Tsukuba Ibaraki Tsukuba Japan

ISBN: (纸本)9798350373332

Clustering is a typical unsupervised learning method for classifying unsupervised data. One of the clustering meth-ods, even-sized clustering based on optimization (ECBO), is a clustering algorithm that imposes a constraint to equalize the size of each cluster. ECBO has been suggested to be effective in delivery and other problems. However, it is limited to Euclidean space. On the other hand, spectral clustering with a wide range of applicability for partitioning graph data has been proposed. In this paper, we propose even-sized spectral clustering, which imposes a size-equal constraint on spectral clustering, and show that it is an extension of the graph partitioning problem. We also verify the validity of the results through numerical examples. © 2024 IEEE.

关键词： Unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

NUMERICAL ANALYSIS FOR CLARIFYING DYNAMIC BUCKLING BEHAVIOR OF PWR-CV SUBJECT TO A SEISMIC EXCITATION 30

NUMERICAL ANALYSIS FOR CLARIFYING DYNAMIC BUCKLING BEHAVIOR ...

引用

30th International Conference on Nuclear engineering, ICONE 2023

作者： Hidaka, Shinjiro Shoji, Gaku Degree Programs in Systems and Information Engineering Graduate School of Science and Technology University of Tsukuba Ibaraki Japan Division of Engineering Mechanics and Energy Institute of Systems and Information Engineering University of Tsukuba Ibaraki Japan

ISBN: (纸本)9784888982566

Clarifying the dynamic buckling behavior of PWR steel reactor containment vessels (PWR-CVs) subject to seismic excitation is critical for improving a numerical seismic design scheme for PWR-CVs. In evaluating the static elastic-plastic buckling strength for PWR-CVs by a finite element (FE) analysis, first we compute the moments and axial forces acting onto the multiple lumped mass beam model of a targeted PWR-CV from the linear dynamic analysis, second using the buckling evaluation formula on JEAG 4601 addenda-1984 we calculate the equivalent static load distributions in terms of shear and axial forces as input loads to the FE model. This static FE buckling analysis for a PWR-CV idealizes dynamic buckling phenomena as be static phenomena, so we cannot basically consider the effects of dynamic behavior on the buckling strength of a PWR-CV. To tackle more realistic buckling assessment of a PWR-CV, we carried out the linear dynamic analysis for a PWR-CV FE numerical model and discussed the features of the response considering the oval vibration modes. It was revealed that the response accompanying the oval vibration modes showed the lower value compared with that from the static FE buckling analysis. In contrast, in a previous study by Hidaka et al., 2022, it was discussed that the buckling load of static FE buckling analysis for a PWR-CV is higher than the buckling load of buckling evaluation formula on JEAG4601. As a result, the ratio of buckling load to response in the linear dynamic FE analysis was within 3% of that in the static FE buckling analysis. © 2023 by JSME.

关键词： Elastoplasticity

来源：评论

学校读者我要写书评

暂无评论

Implementation and Performance Evaluation of Memory System Using Addressable Cache for HPC Applications on HBM2 Equipped FPGAs 28th

Implementation and Performance Evaluation of Memory System...

引用

28th International European Conference on Parallel and Distributed Computing , Euro-Par 2022

作者： Fujita, Norihisa Kobayashi, Ryohei Yamaguchi, Yoshiki Boku, Taisuke Center for Computational Sciences University of Tsukuba Tsukuba Japan Degree Programs in Systems and Information Engineering Graduate School of Science and Technology University of Tsukuba Tsukuba Japan

ISBN: (纸本)9783031312083

When we apply field programmable gate arrays (FPGAs) as HPC accelerators, their memory bandwidth presents a significant challenge because it is not comparable to those of other HPC accelerators. In this paper, we propose a memory system for HBM2-equipped FPGAs and HPC applications that uses block RAMs as an addressable cache implemented between HBM2 and an application. This architecture enables data transfer between HBM2 and the cache bulk and allows an application to utilize fast random access on BRAMs. This study demonstrates the implementation and performance evaluation of our new memory system for HPC and HBM2 on an FPGA. Furthermore, we describe the API that can be used to control this system from the host. We implement RISC-V cores in an FPGA as controllers to realize fine-grain data transfer control and to prevent overheads derived from the PCI Express bus. The proposed system is implemented on eight memory channels and achieves 102.7 GB/s of the bandwidth. It overcomes the memory bandwidth of conventional FPGA boards with four channels of DDR4 memory despite using only 8 of 32 channels of the HBM2. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Data transfer

来源：评论

学校读者我要写书评

暂无评论

EFFECTS OF REFRACTIVE INDEX AND ASPECT RATIO ON THE PARTICLE IMAGES IN A LEVITATED DROPLET

引用

Multiphase science and technology 2024年第1期36卷 13-26页

作者： Gatete, Eugene Kaneko, Akiko Shen, Biao University of Tsukuba Systems and Information Engineering Graduate School of Science and Technology Degree Program of Engineering Mechanics and Energy 1-1-1 Tennodai Tsukuba Ibaraki305-8573 Japan University of Tsukuba Faculty of Engineering of Information Engineering and Systems 1-1-1 Tennodai Tsukuba Ibaraki305-8573 Japan

A technique using droplets suspended by ultrasound has attracted attention as one of the containerless processing methods. While this can avoid contamination from the container, it is known that ultrasonic levitation creates flow fields inside and outside the droplet. For more precise droplet control, it is desirable to elucidate the internal flow of the droplet, and measurements of the internal flow have been performed using the particle image velocimetry (PIV). The aim of this study is to elucidate the internal flow field behavior by solving optical problems and improving the accuracy of velocity field measurements in levitated droplets. The fluid that is to be investigated is scattered with small tracer particles and illuminated by a laser to capture the flow on a created laser sheet. The curvature distortion is successfully visualized using the PIV approach toward distortion correction using calibration methods. The curvature distortion illustration was performed based on the refractive index and aspect ratio of the simulated droplet in acrylic materials. The fluid flow, affected by droplet curvature and refractive index, has been visualized for both levitated and simulated droplets. The experimental results showed that the droplet curvature can be distorted in two types such as radial and tangential distortions and increase as the refractive index and aspect ratio increases. © 2024 by Begell House, Inc.

关键词： Refractive index

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

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

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：