检索结果-内蒙古大学图书馆

DeepCNN:Spectro-temporal feature representation for speech emotion recognition

CAai Transactions on Intelligence Technology 2023年第2期8卷 401-417页

作者： Nasir Saleem Jiechao Gao Rizwana Irfan Ahmad Almadhor Hafiz Tayyab Rauf Yudong Zhang Seifedine Kadry Department of Electrical Engineering Faculty of Engineering and TechnologyGomal UniversityD.I.KhanPakistan Department of Computer Science University of VirginiaCharlottesvilleVirginiaUSA Department of Information Technology College of Computing and Information Technology at KhulaisUniversity of JeddahJeddahSaudi Arabia Department of Computer Engineering and Networks College of Computer and Information SciencesJouf UniversitySkaka AljoufSaudi Arabia Centre for Smart Systems AI and CybersecurityStaffordshire UniversityStoke‐on‐TrentUK School of Computing and Mathematical Sciences University of LeicesterLeicesterUK Department of Applied Data Science Noroff University CollegeKristiansandNorway Artificial Intelligence Research Center(AIRC) Ajman UniversityAjmanUnited Arab Emirates Department of Electrical and Computer Engineering Lebanese American UniversityByblosLebanon

Speech emotion recognition(SER)is an important research problem in human-computer interaction *** representation and extraction of features are significant challenges in SER *** the promising results of recent studies,they generally do not leverage progressive fusion techniques for effective feature representation and increasing receptive *** mitigate this problem,this article proposes DeepCNN,which is a fusion of spectral and temporal features of emotional speech by parallelising convolutional neural networks(CNNs)and a convolution layer-based *** parallel CNNs are applied to extract the spectral features(2D-CNN)and temporal features(1D-CNN)representations.A 2D-convolution layer-based transformer module extracts spectro-temporal features and concatenates them with features from parallel *** learnt low-level concatenated features are then applied to a deep framework of convolutional blocks,which retrieves high-level feature representation and subsequently categorises the emotional states using an attention gated recurrent unit and classification *** fusion technique results in a deeper hierarchical feature representation at a lower computational cost while simultaneously expanding the filter depth and reducing the feature *** Berlin Database of Emotional Speech(EMO-BD)and Interactive Emotional Dyadic Motion Capture(IEMOCAP)datasets are used in experiments to recognise distinct speech *** efficient spectral and temporal feature representation,the proposed SER model achieves 94.2%accuracy for different emotions on the EMO-BD and 81.1%accuracy on the IEMOCAP dataset *** proposed SER system,DeepCNN,outperforms the baseline SER systems in terms of emotion recognition accuracy on the EMO-BD and IEMOCAP datasets.

关键词： decision making deep learning

来源：评论

学校读者我要写书评

暂无评论

A Dual-Prompting for Interpretable Mental Health Language Models 9

A Dual-Prompting for Interpretable Mental Health Language Mo...

引用

9th Workshop on Computational Linguistics and Clinical Psychology, CLPsych 2024

作者： Jeon, Hyolim Yoo, Dongje Lee, Daeun Son, Sejung Kim, Seungbae Han, Jinyoung Department of Applied Artificial Intelligence Sungkyunkwan University Seoul Korea Republic of Department of Human-AI Interaction Sungkyunkwan University Seoul Korea Republic of Department of Computer Engineering Chung-Ang University Seoul Korea Republic of Computer Science & Engineering Department University of South Florida TampaFL United States

ISBN: (纸本)9798891760806

Despite the increasing demand for ai-based mental health monitoring tools, their practical utility for clinicians is limited by the lack of interpretability. The CLPsych 2024 Shared Task1 aims to enhance the interpretability of Large Language Models (LLMs), particularly in mental health analysis, by providing evidence of suicidality through linguistic content. We propose a dual-prompting approach: (i) Knowledge-aware evidence extraction by leveraging the expert identity and a suicide dictionary with a mental health-specific LLM;and (ii) Evidence summarization by employing an LLM-based consistency evaluator. Comprehensive experiments demonstrate the effectiveness of combining domain-specific information, revealing performance improvements and the approach’s potential to aid clinicians in assessing mental state progression. ©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

Slice-Level Label Attention with Global-Guided Attention Regularization for Multi-Label Classification in Knee MRI Sequences

Slice-Level Label Attention with Global-Guided Attention Reg...

引用

2024 IEEE International Conference on Big Data, BigData 2024

作者： Yang, Jingzhi Shen, Hua Liu, Ziyu Wu, Weilong Wu, Ji Ni, Ming Yuan, Huishu Fu, Xiangling Li, Miao National Pilot Software Engineering School Bupt School of Computer Science Beijing China Tsinghua University College of Ai Department of Electronic Engineering Beijing China Peking University Third Hospital Beijing China

ISBN: (纸本)9798350362480

Magnetic Resonance Imaging (MRI) is crucial for diagnosing various knee-related diseases, and developing automatic diagnostic models based on knee MRI data is highly valuable. However, this task presents significant challenges due to the need to manage MRI data with multiple sequences and numerous images, where different diseases are often associated with specific images within certain sequences. To address these challenges, we propose a multi-label classification framework designed to effectively process MRI data and handle a large-scale label space encompassing hundreds of disease categories. Our approach introduces a Slice-Level Label Attention mechanism, which enables the model to learn the alignment between labels and images within sequences, thereby enhancing both performance and interpretability. Additionally, we present a Global-Guided Attention Regularization mechanism that further improves the consistency and robustness of the Slice-Level Label Attention results. We validate our framework on a large-scale MRI dataset involving multi-label classification across hundreds of fine-grained disease categories. Experimental results demonstrate that our method not only achieves superior performance but also provides more robust and consistent interpretability. © 2024 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Predictive Analysis Using Modified Inception Net and Support Vector Machine for Forecasting Copra Yield in Agriculture 3

Predictive Analysis Using Modified Inception Net and Support...

引用

3rd IEEE International Conference on Distributed Computing and Electrical Circuits and Electronics, ICDCECE 2024

作者： Kavitha, M.G. Raviprakash, M.L. Al-Attabi, Kassem Sathyavani, B. Sundaram, S. Meenakshi Kalpataru Institute Of Technology Department Of Computer Science And Engineering Tiptur India College Of Technical Engineering The Islamic University Department Of Computers Techniques Engineering Najaf Iraq Sr University Department Of Electrical And Electronics Engineering Warangal India Nitte Meenakshi Institute Of Technology Department Of Ai & Ds Bengaluru India

ISBN: (纸本)9798350318609

One of the major economic sectors in various states is copra agriculture, with a significant portion of its global production coming from countries such as India, Indonesia, and Sri Lanka. However, the severity of copra issues can vary depending on factors such as moisture levels, fungal presence, and the quality of the copra, leading to imbalanced data and potential overfitting in copra image analysis. Threshold-based segmentation is involved to identify different segments of copra based on factors like moisture content. Higher moisture content may indicate areas prone to fungal growth or poor drying conditions, while segments with uniform colour and texture may signify high-quality copra. The proposed approach involves using the Modified Inception Net (MIN) for feature extraction and Support Vector Machine (SVM) for hyperparameter tuning. SVM handles the kernel function in nonlinear and high-dimensional spaces, enhancing its capabilities and mitigating overfitting. However, the MIN-SVM combination may require significant computational resources, particularly during training and parameter tuning. The proposed MIN-SVM method achieves promising performance metrics, including a Mean Absolute Error (MAE) of 0.071, Mean Square Error (MSE) of 0.0125, Matthews Correlation Coefficient (MCC) of 0.620 and R2 value of 0.9410. This performance is compared to existing methods such as Back Propagation Algorithm (BPA), K-means and SVM. © 2024 IEEE.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

Noise contrastive alignment of language models with explicit rewards 24

Noise contrastive alignment of language models with explicit...

引用

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Huayu Chen Guande He Lifan Yuan Ganqu Cui Hang Su Jun Zhu Department of Computer Science and Technology Tsinghua University and Institute for AI BNRist Center Tsinghua-Bosch Joint ML Center THBI Lab Tsinghua University Department of Computer Science and Technology Tsinghua University Department of Computer Science and Technology Tsinghua University and Institute for AI BNRist Center Tsinghua-Bosch Joint ML Center THBI Lab Tsinghua University and Zhongguancun Laboratory Beijing China

ISBN: (纸本)9798331314385

User intentions are typically formalized as evaluation rewards to be maximized when fine-tuning language models (LMs). Existing alignment methods, such as Direct Preference Optimization (DPO), are mainly tailored for pairwise preference data where rewards are implicitly defined rather than explicitly given. In this paper, we introduce a general framework for LM alignment, leveraging Noise Contrastive Estimation (NCE) to bridge the gap in handling reward datasets explicitly annotated with scalar evaluations. Our framework comprises two parallel algorithms, NCA and InfoNCA, both enabling the direct extraction of an LM policy from reward data as well as preference data. Notably, we show that the DPO loss is a special case of our proposed InfoNCA objective under pairwise preference settings, thereby integrating and extending current alignment theories. By comparing NCA and InfoNCA, we demonstrate that the well-observed decreasing-likelihood trend of DPO/InfoNCA is caused by their focus on adjusting relative likelihood across different responses. In contrast, NCA optimizes the absolute likelihood for each response, thereby effectively preventing the chosen likelihood from decreasing. We evaluate our methods in both reward and preference settings with Mistral-8 x7B and 7B models. Experiments suggest that InfoNCA/NCA surpasses various preference baselines when reward datasets are available. We also find NCA significantly outperforms DPO in complex reasoning tasks like math and coding. Code: https://***/thu-ml/Noise-Contrastive-Alignment.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Fast Vocabulary Transfer for Language Model Compression

Fast Vocabulary Transfer for Language Model Compression

引用

2022 Conference on Empirical Methods in Natural Language Processing, EMNLP 2022

作者： Gee, Leonidas Zugarini, Andrea Rigutini, Leonardo Torroni, Paolo Expert.ai Italy University of Siena Department of Computer Science and Engineering University of Bologna

ISBN: (纸本)9781952148255

Real-world business applications require a trade-off between language model performance and size. We propose a new method for model compression that relies on vocabulary transfer. We evaluate the method on various vertical domains and downstream tasks. Our results indicate that vocabulary transfer can be effectively used in combination with other compression techniques, yielding a significant reduction in model size and inference time while marginally compromising on performance. © 2022 Association for Computational Linguistics.

关键词： Economic and social effects

来源：评论

学校读者我要写书评

暂无评论

An Effective Tensor Regression with Latent Sparse Regularization

引用

Journal of Data science 2022年第2期20卷 228-252页

作者： Chen, Ko-Shin Xu, Tingyang Liang, Guannan Tong, Qianqian Song, Minghu Bi, Jinbo Department of Computer Science and Engineering University of Connecticut United States Tencent AI Lab China Department of Biomedical Engineering University of Connecticut United States

As data acquisition technologies advance, longitudinal analysis is facing challenges of exploring complex feature patterns from high-dimensional data and modeling potential temporally lagged effects of features on a response. We propose a tensor-based model to analyze multidimensional data. It simultaneously discovers patterns in features and reveals whether features observed at past time points have impact on current outcomes. The model coefficient, a k-mode tensor, is decomposed into a summation of k tensors of the same dimension. We introduce a so-called latent F-1 norm that can be applied to the coefficient tensor to performed structured selection of features. Specifically, features will be selected along each mode of the tensor. The proposed model takes into account within-subject correlations by employing a tensor-based quadratic inference function. An asymptotic analysis shows that our model can identify true support when the sample size approaches to infinity. To solve the corresponding optimization problem, we develop a linearized block coordinate descent algorithm and prove its convergence for a fixed sample size. Computational results on synthetic datasets and real-life fMRI and EEG datasets demonstrate the superior performance of the proposed approach over existing techniques. © 2022 The Author(s).

关键词： longitudinal data quadratic inference function tensors

来源：评论

学校读者我要写书评

暂无评论

Adaptive Label Smoothing with Self-Knowledge in Natural Language Generation

Adaptive Label Smoothing with Self-Knowledge in Natural Lang...

引用

2022 Conference on Empirical Methods in Natural Language Processing, EMNLP 2022

作者： Lee, Dongkyu Cheung, Ka Chun Zhang, Nevin L. Department of Computer Science and Engineering HKUST Hong Kong NVIDIA AI Technology Center NVIDIA

Overconfidence has been shown to impair generalization and calibration of a neural network. Previous studies remedy this issue by adding a regularization term to a loss function, preventing a model from making a peaked distribution. Label smoothing smoothes target labels with a pre-defined prior label distribution;as a result, a model is learned to maximize the likelihood of predicting the soft label. Nonetheless, the amount of smoothing is the same in all samples and remains fixed in training. In other words, label smoothing does not reflect the change in probability distribution mapped by a model over the course of training. To address this issue, we propose a regularization scheme that brings dynamic nature into the smoothing parameter by taking model probability distribution into account, thereby varying the parameter per instance. A model in training self-regulates the extent of smoothing on the fly during forward propagation. Furthermore, inspired by recent work in bridging label smoothing and knowledge distillation, our work utilizes self-knowledge as a prior label distribution in softening target labels, and presents theoretical support for the regularization effect by knowledge distillation and the dynamic smoothing parameter. Our regularizer is validated comprehensively, and the result illustrates marked improvements in model generalization and calibration, enhancing robustness and trustworthiness of a model. © 2022 Association for Computational Linguistics.

关键词： Distillation

来源：评论

学校读者我要写书评

暂无评论

Context is King: Rethinking Query Expansion in Modern Search Engines

Context is King: Rethinking Query Expansion in Modern Search...

引用

Artificial Intelligence and Data engineering (aiDE), International Conference on

作者： Krishnan Batri S Lakshmi S Ramesh N Sangeetha S. Annamalai School of Computer Science and Engineering Jain(Deemed-to-be-University) Bengaluru India Electronics and Communication Engineering Jain(Deemed-to-be-University) Bengaluru India Department of AI&ML City Engineering College Bengaluru India

ISBN: (数字)9798331527518

ISBN: (纸本)9798331527525

In the era of digital information overload, search engines have evolved beyond simple keyword matching to embrace context-aware methodologies. However, the task of accurately interpreting user intent remains a critical challenge, with traditional query expansion techniques often falling short of capturing semantic and contextual nuances. This paper critically examines the evolution of query expansion strategies, highlighting their limitations and exploring state-of-the-art techniques such as machine learning models, knowledge graphs, and transformer-based systems like BERT and GPT. Experimental results compare traditional methods like synonym expansion with modern approaches, demonstrating the superior relevance and diversity of results from context-aware systems. The study underscores the importance of integrating real-time contextual signals and dynamic user preferences to enhance search engine performance, paving the way for more personalized and accurate retrieval systems.

关键词： Adaptation models Quantum computing Biological system modeling Semantics Machine learning Knowledge graphs Search engines Transformers Real-time systems Natural language processing

来源：评论

学校读者我要写书评

暂无评论

Enhancing Supply Chain Regulatory Compliance with Blockchain and Lasso Regression Integration 2

Enhancing Supply Chain Regulatory Compliance with Blockchain...

引用

2nd IEEE International Conference on IoT, Communication and Automation Technology, ICICAT 2024

作者： Aelgani, Vivekanand Jothi, E. Hussein, Ramy Riad Ramesh, Banoth Sharma, Himanshu Uganya, G. CMR College of Engineering & Technology Department of CSE Telangana Hyderabad India New Prince Shri Bhavani College of Engineering and Technology Department of IT Tamil Nadu Chennai600073 India The Islamic University College of Technical Engineering Department of Computers Techniques Engineering Najaf Iraq GRIET Bachupally Department of AI&ML Telangana Hyderabad India GLA University Department of Computer Engineering & Applications Mathura India Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology Department of ECE Chennai600062 India

ISBN: (纸本)9798350368109

Adopting lasso regression besides blockchain gives the best solution in improving the regulatory compliance in supply chains. For this integration, the feature of blockchain technology which applies for secure, transparency and unchangeable record keeping is used, For lasso regression the variable selection and predictive model are used. Thus, blockchain offers supply chain participants an opportunity to perform all transactions in a decentralized manner and maintain an unalterable, open record of all the transactions that occur in the supply chain. At the same time, the lasso regression models are used to predict compliance risks based on the data obtained and choosing the features that might be useful for this analysis from a vast set of possibilities. Smart contracts are the other strategies used in performing compliance checks in real-time and reporting instances that indicate non-compliance so that the necessary intervention is done. This system does not only administers and optimizes the press compliance procedure but also improves prediction rate and reduces the risk of encountering new compliance issues. Further, the changes are made continuously in the model and the feedbacks from the stakeholders add to the fact that the system is fit to adapt to the changes in the regulatory environment and supply chain industry. In doing so, it also enhances the visibility, relevancy, reliability, and effectiveness of the regulation compliance enhancing better trust and operation in the overall supply chain. © 2024 IEEE.

关键词： Smart contract

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：