检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Raissi, Tina Lüscher, Christoph Gunz, Moritz Schlüter, Ralf Ney, Hermann Machine Learning and Human Language Technology Computer Science Department RWTH Aachen University Aachen52074 Germany AppTek GmbH Aachen52062 Germany

Building competitive hybrid hidden Markov model (HMM) systems for automatic speech recognition (ASR) requires a complex multi-stage pipeline consisting of several training criteria. The recent sequence-to-sequence models offer the advantage of having simpler pipelines that can start from-scratch. We propose a purely neural based single-stage from-scratch pipeline for a context-dependent hybrid HMM that offers similar simplicity. We use an alignment from a full-sum trained zero-order posterior HMM with a BLSTM encoder. We show that with this alignment we can build a Conformer factored hybrid that performs even better than both a state-of-the-art classic hybrid and a factored hybrid trained with alignments taken from more complex Gaussian mixture based systems. Our finding is confirmed on Switchboard 300h and LibriSpeech 960h tasks with comparable results to other approaches in the literature, and by additionally relying on a responsible choice of available computational resources. © 2023, CC BY.

关键词： Hidden Markov models

来源：评论

学校读者我要写书评

暂无评论

INVESTIGATING THE EFFECT OF LANGUAGE MODELS IN SEQUENCE DISCRIMINATIVE TRAINING FOR NEURAL TRANSDUCERS

arXiv

引用

arXiv 2023年

作者： Yang, Zijian Zhou, Wei Schlüter, Ralf Ney, Hermann Machine Learning and Human Language Technology Computer Science Department RWTH Aachen University Aachen52074 Germany AppTek GmbH Aachen52062 Germany

In this work, we investigate the effect of language models (LMs) with different context lengths and label units (phoneme vs. word) used in sequence discriminative training for phoneme-based neural transducers. Both lattice-free and N-best-list approaches are examined. For lattice-free methods with phoneme-level LMs, we propose a method to approximate the context history to employ LMs with full-context dependency. This approximation can be extended to arbitrary context length and enables the usage of word-level LMs in lattice-free methods. Moreover, a systematic comparison is conducted across lattice-free and N-best-list-based methods. Experimental results on Librispeech show that using the word-level LM in training outperforms the phoneme-level LM. Besides, we find that the context size of the LM used for probability computation has a limited effect on performance. Moreover, our results reveal the pivotal importance of the hypothesis space quality in sequence discriminative training. Copyright © 2023, The Authors. All rights reserved.

关键词： Transducers

来源：评论

学校读者我要写书评

暂无评论

Data Augmentation for Classifying Multiple Sclerosis Severity through Inertial Measurement Unit-Based Gait Analysis

Data Augmentation for Classifying Multiple Sclerosis Severit...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Jessica B. Li John W. Farrell Damian Valles Department of Statistics Department of Computer Science Northwestern University Evanston USA Department of Health and Human Performance Texas State University San Marcos USA Ingram School of Engineering Texas State University San Marcos USA

ISBN: (数字)9798350386226

ISBN: (纸本)9798350386233

Gait impairments are highly prevalent in persons with multiple sclerosis (PwMS), contributing to difficulties in daily activities. Disability is commonly assessed using the Expanded Disability Status Scale (EDSS). However, the EDSS lacks detailed gait analysis for assessing the severity of multiple sclerosis. To address this, gait analysis tools such as inertial measurement units are commonly used to understand walking patterns in PwMS. Another concern is that collecting sufficient gait data becomes challenging due to limited participation in studies. This research acknowledges these limitations and proposes using a variational autoencoder to address this issue. Additionally, the study explores the feasibility of classification models aimed at assisting the quantification of disability of PwMS based on individuals’ gait patterns.

关键词： Legged locomotion Multiple sclerosis Measurement units Autoencoders Pulse width modulation Gain measurement Predictive models Biomedical measurement Data augmentation Data models

来源：评论

学校读者我要写书评

暂无评论

Prompting Large Vision-Language Models for Compositional Reasoning

arXiv

引用

arXiv 2024年

作者： Ossowski, Timothy Jiang, Ming Hu, Junjie Department of Computer Science University of Wisconsin MadisonWI United States Department of Biostatistics and Medical Informatics University of Wisconsin MadisonWI United States Department of Human-centered Computing Indiana University IndianapolisIN United States

Vision-language models such as CLIP have shown impressive capabilities in encoding texts and images into aligned embeddings, enabling the retrieval of multimodal data in a shared embedding space. However, these embedding-based models still face challenges in effectively matching images and texts with similar visio-linguistic compositionality, as evidenced by their performance on the recent Winoground dataset. In this paper, we argue that this limitation stems from two factors: the use of single vector representations for complex multimodal data, and the absence of step-by-step reasoning in these embedding-based methods. To address this issue, we make an exploratory step using a novel generative method that prompts large vision-language models (e.g., GPT-4) to depict images and perform compositional reasoning. Our method outperforms other embedding-based methods on the Winoground dataset, and obtains further improvement of up to 10% accuracy when enhanced with the optimal description. More importantly, we provide a fine-grained error analysis of our method's outputs, highlighting the key bottleneck in understanding image contents by existing VLMs. Copyright © 2024, The Authors. All rights reserved.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

An Observational Study Investigating Verbal Communication Differences in Online vs. In-person Group Exercise Training 66th

An Observational Study Investigating Verbal Communication Di...

引用

66th International Annual Meeting of the human Factors and Ergonomics Society, HFES 2022

作者： Islam, Md Shafiqul Sayani, Naima Harden, Samantha M. Won Lee, Sang Lim, Sol Department of Industrial and Systems Engineering Virginia Tech VA United States School of Neuroscience Virginia Tech VA United States Department of Human Nutrition Foods and Exercise Virginia Tech VA United States Department of Computer Science Virginia Tech VA United States

来源：评论

学校读者我要写书评

暂无评论

Mechanical Design with Experimental Verification of a Lightweight Exoskeleton Chair

引用

Journal of Bionic Engineering 2021年第2期18卷 319-332页

作者： Zihao Du Zefeng Yan Tiantian Huang Ou Bai Qin Huang Bin Han State Key Laboratory of Digital Manufacturing Equipment and Technology Huazhong University of Science and TechnologyWuhan 430074China Human Cyber-Physical Systems Laboratory Department of Electrical and Computer EngineeringFlorida International UniversityMiami 33174USA Department of Rehabilitation Medicine Union HospitalTongji Medical CollegeHuazhong University of Science and TechnologyWuhan 430022China

In this study,a human-chair model was developed as the basis for a wearable-chair design.A prototype chair,HUST-EC,based on the model was fabricated and *** the optimization under the golden divisional method,an optimized simulation of the operating mode with the lowest chair height was implemented.A novel multi-link support structure has been established with parameters optimized using Matlab *** stress analysis of the solid models was conducted to ensure the adequate support from the designed chair for the *** subjects participated in the evaluation experiment,who performed both static tasks and dynamic *** experimental results consisted of subjective evaluation and objective *** experimental data demonstrate that(1)the HUST-EC can effectively reduce the activation level of related muscles at a variety of tasks;(2)the plantar pressure was reduced by 54%–67%;(3)the angle between the upper body and the vertical axis was reduced by 59%–77%;(4)the subjective scores for chair comfortability,portability,and stability were all higher than *** results further revealed that the designed chair can reduce the musculoskeletal burden and may improve work efficiency.

关键词： exoskeleton wearable chair bionic robot mechanism design ergonomics

来源：评论

学校读者我要写书评

暂无评论

AB1377 FULLY AUTOMATIC ASSESSMENT OF NAIL FOLD CAPILLAROSCOPY SOFTWARE – PILOT STUDY

引用

Annals of the Rheumatic Diseases 2022年 81卷 1793-1794页

作者： O. Brzezińska K. Rychlicki-Kicior J. Makowska Medical University of Lodz Department of Rheumatology Łódź Poland University of Economics and Human Sciences in Warsaw Computer Science Warsaw Poland

Background The nail fold capillaroscopy is one of the basic techniques used in diagnosis and monitoring the course of connective tissue diseases, primarily a systemic sclerosis. However, the assessment of the capillary image is time-consuming and subjective, this makes it difficult for a detailed comparison of studies assessed by various physicians. Objectives The aim of this study was to validate an automated software for classification the nail fold capillaroscopy as normal or pathological and counting the numerous of vessels on a millimetre. Methods 100 correct and 100 pathological images were selected from the capillaroscopic image database. The original examination was performed with Dinolite MEDL4N Pro. The classic, manual evaluation of the capillaroscopy was performed twice. The imagines was classified as correct or pathological and mean number of capillaries per millimetre was calculated. The photos were then exported to the training program created for the study, the region of interest (ROI) and individual capillaries were marked. The calculation of the number of vessels was made at a magnification of about 50x (ensuring the maximum quality of the image obtained during a classic examination). The neural network was trained using the *** library (based on PyTorch). The ResNet-34 deep residual neural network was chosen, 10-fold cross-validation with validation and test set was performed, using Darknet-YoloV3 state of the art neural network in a GPU-optimized (P5000 GPU) environment. For calculation of 1mm capillaries, additional detection mechanism was designed, in order to automatically detect the scale of the image and transform it into proper pixel dimensions. Results The results obtained under neural network training have been referred to the results obtained as part of a manual photo assessment. For the image classification correct vs pathological test sensitivity 89.0% and specificity 86.9% was obtained. For validation (20% of images have been drawn

关键词：

来源：评论

学校读者我要写书评

暂无评论

On the potential for human-centered, cognitively inspired AI to bridge the gap between optimism and reality for autonomous robotics in healthcare: a respectful critique

引用

Proceedings of the International Symposium on human Factors and Ergonomics in Health Care 2024年第1期13卷 106 - 112页

作者： Stephanie Tulk Jesso Christopher Greene Shiqi Zhang Amy Booth Michael DiFabio Grace Babalola Adeola Adegbemijo Sumantra Sarkar 1Department of Systems Science and Industrial Engineering SUNY Binghamton2Human-Centered Mindful Technologies Lab SUNY Binghamton 1Department of Systems Science and Industrial Engineering SUNY Binghamton 3Department of Computer Science SUNY Binghamton 4New York United Health Services 5School of Management SUNY Binghamton

Historically, the fields of computer science, cognitive science, and neuroscience have been tightly linked. To date, this collaboration has yielded major advances in how the brain and mind are understood, as well as the ways in which artificial minds can be constructed to serve as new collaborators to humans. Yet there are still significant gaps between the capabilities of state-of-the-art autonomous robots and the expectations developed by real users who are now encountering autonomous robots on the job. We present our views as well as a case study of our evaluation of two autonomous robots intended to aid nurses within hospital settings: Moxi and TUG. Both cobots were originally considered for procurement by our collaborating healthcare system,UHS, at which time our team began the process of trying to systematically vet each option to aid in the decision making process. What we found was a lack of evidence on either platform in academic literature, which led us to analyze user comments on social media. In order to improve the fit of autonomous robots into human environments, research must be conducted and evidence must be shared, and the human Factors community can aid in this effort

关键词： Cognitively Plausible AI Autonomous Robots Cobots Healthcare Nursing 4E Cognition

来源：评论

学校读者我要写书评

暂无评论

Deep Learning-Based Operational Risk Prevention and Control Monitoring of Smart Financial Systems 2

Deep Learning-Based Operational Risk Prevention and Control ...

引用

2nd IEEE International Conference on Advances in Computing, Communication and Applied Informatics, ACCAI 2023

作者： Jain, Sonal Hemalatha, Raavi Mubeen, Samiya Jethy, Jayashree Kandewar, Priti Pujari, Rajendra Kiit University School of Social Financial & Human Sciences Bhubaneswar India Vishnu Institute of Technology Department of Computer Science and Engineering Bhimavaram India Iibs Department of Management Studies Bengaluru India Rama Devi Women's University Department of Commerce Bhubaneswar India St. Martin's Engineering College Department of Cse India Institute of Management and Rural Development Administration Sangli India

ISBN: (纸本)9798350315905

The risk of carrying on with work has expanded throughout recent years as the monetary business has developed consistently. Significant risk cases are normal, the circumstances turning out to be more convoluted, and the method for carrying out violations are tucked away among them. The pretreatment of interior and outer monetary information as well as the construction plan of intermittent NNs are the key review subjects of this work. Its goal is to foster a deep learning NN-based model for controlling monetary risk, eventually bringing down risk. The example information are first preprocessed utilizing the Fringe Destroyed calculation, the lopsidedness in the information is eliminated utilizing the oversampling approach, lastly, the example information with time series attributes are handled utilizing the long momentary memory deep NNs calculation. The end explore exhibits that, when contrasted with traditional methodologies, LSTM has a higher precision, coming to 1.9715;the example readiness technique and risk control model recommended in this study have a better capacity than recognize false customers;and the actual model has a quicker cycle productivity. The persistent control of perils coming from human exercises, interior cycles, frameworks, and outside occasions is known as operational risk the board (ORM). © 2023 IEEE.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Evolving Deep Neural Network by Customized Moth Flame Optimization Algorithm for Underwater Targets Recognition

arXiv

引用

arXiv 2023年

作者： Khishe, Mohammad Mohammadi, Mokhtar Rashid, Tarik A. Mahmud, Hoger Mirjalili, Seyedali Department of Marine Electronics and Communication Engineering Imam Khomeini Marine Science University Nowshahr Iran Department of Information Technology College of Engineering and Computer Science Lebanese French University KR Iraq Computer Science and Engineering Department Science and Engineering Science University of Kurdistan Hewler KR Erbil Iraq Department of Computer Science College of Science and Technology University of Human Development KR Sulaymaniyah Iraq Centre for Artificial Intelligence Research and Optimization Torrens University Australia Yonsei Frontier Lab Yonsei University Seoul Korea Republic of

This chapter proposes using the Moth Flame Optimization (MFO) algorithm for fine-tuning a Deep Neural Network to recognize different underwater sonar datasets. Same as other models evolved by metaheuristic algorithms, premature convergence, trapping in local minima, and failure to converge in a reasonable time are three defects MFO confronts in solving problems with high-dimension search space. Spiral flying is the key component of the MFO as it determines how the moths adjust their positions in relation to flames;thereby, the shape of spiral motions can regulate the transition behavior between the exploration and exploitation phases. Therefore, this chapter investigates the efficiency of seven spiral motions with different curvatures and slopes in the performance of the MFO, especially for underwater target classification tasks. To assess the performance of the customized model, in addition to benchmark Sejnowski & Gorman's dataset, two experimental sonar datasets, i.e., the passive sonar and active datasets, are exploited. The results of MFO and its modifications are compared to four novel nature-inspired algorithms, including Heap-Based Optimizer (HBO), Chimp Optimization Algorithm (ChOA), Ant Lion Optimization (ALO), Stochastic Fractals Search (SFS), as well as the classic Particle Swarm Optimization (PSO). The results confirm that the customized MFO shows better performance than the other state-of-the-art models so that the classification rates are increased 1.5979, 0.9985, and 2.0879 for Sejnowski & Gorman, passive, and active datasets, respectively. The results also approve that time complexity is not significantly increased by using different spiral motions. © 2023, CC BY.

关键词： Deep neural networks

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：