检索结果-内蒙古大学图书馆

A review of explainable AI in medical imaging: implications and applications

International Journal of computers and Applications 2024年第11期46卷 983-997页

作者： Kinger, Shakti Kulkarni, Vrushali Department of Computer Engineering and Technology WPU School of Computer Science & Engineering Dr. Vishwanath Karad MIT World Peace University Maharashtra India

Deep learning approaches have attained remarkable success across various artificial intelligence applications, spanning healthcare, finance, and autonomous vehicles, profoundly impacting human existence. However, their black-box nature, lack of transparency, and inability to elucidate conclusions have hindered their adoption in high-risk applications. Explainable Artificial Intelligence (XAI) encompasses a suite of tools, approaches, and algorithms aimed at furnishing highly accurate explanations while preserving robust accuracy. This research paper investigates the intricate and evolving domain of Explainable Artificial Intelligence (XAI), particularly its implications in healthcare, with the objective of fostering trust in AI utilization within this domain. It encompasses a spectrum of topics, including a compendium of tasks XAI should fulfil in medical imaging, an examination of current methodologies for yielding transparent and understandable outcomes in medical imaging, criteria for assessing AI system explainability, and recommendations for integrating XAI into medical imaging practices. This review facilitates the selection of suitable and efficient XAI techniques for medical imaging while aiding developers in grasping the fundamentals of these methodologies. © 2024 Informa UK Limited, trading as Taylor & Francis Group.

关键词： Medical imaging

来源：评论

学校读者我要写书评

暂无评论

High-accuracy sampling from constrained spaces with the Metropolis-adjusted Preconditioned Langevin Algorithm 36

High-accuracy sampling from constrained spaces with the Metr...

引用

36th International Conference on Algorithmic Learning Theory, ALT 2025

作者： Srinivasan, Vishwak Wibisono, Andre Wilson, Ashia Department of Electrical Engineering and Computer Science MIT United States Department of Computer Science Yale University United States

We propose a first-order sampling method called the Metropolis-adjusted Preconditioned Langevin Algorithm for approximate sampling from a target distribution whose support is a proper convex subset of Rd. Our proposed method is the result of applying a Metropolis-Hastings filter to the Markov chain formed by a single step of the preconditioned Langevin algorithm with a metric G, and is motivated by the natural gradient descent algorithm for optimisation. We derive non-asymptotic upper bounds for the mixing time of this method for sampling from target distributions whose potentials are bounded relative to G, and for exponential distributions restricted to the support. Our analysis suggests that if G satisfies stronger notions of self-concordance introduced in Kook and Vempala (2024), then these mixing time upper bounds have a strictly better dependence on the dimension than when G is merely self-concordant. Our method is a high-accuracy sampler due to the polylogarithmic dependence on the error tolerance in our mixing time upper bounds. © 2025 V. Srinivasan, A. Wibisono & A. Wilson.

关键词： Markov chains

来源：评论

学校读者我要写书评

暂无评论

Enhancing coordinated target tracking: innovative particle filters with meta-heuristic integration and advanced model validation

引用

Soft Computing 2025年第7期29卷 1-32页

作者： Das, Nitish Kulkarni, Nilima Department of Computer Science and Engineering MIT Art Design and Technology University Loni Kalbhor Pune412201 India

Coordinated target tracking and its augmented variant represent significant challenges in modern surveillance. These tasks are essential for accurately localizing and predicting the movement of dynamic targets. This article introduces an advanced particle filter algorithm with a packet dropout mechanism to enhance tracking accuracy. Additionally, it proposes three meta-heuristic-driven particle filters: the Particle Filter integrated with the Chaotic Gravitational Search Algorithm (PF-CGSA), the Particle Filter integrated with the Improved Grey Wolf Optimizer (PF-IGWO), and the Particle Filter integrated with the Imperialist Competitive Algorithm with Regional Domination Policy (PF-ICARD). These methods are tested using three sophisticated models: the Coordinated Turn Target Tracking Model (CT model), the Augmented Coordinated Turn Target Tracking Model with Cartesian Velocity (ACT1 model), and the Augmented Coordinated Turn Target Tracking Model with Polar Velocity (ACT2 model). Experimental results demonstrate that PF-CGSA, PF-IGWO, and PF-ICARD achieve higher convergence speeds and accuracy. They also effectively mitigate premature convergence phenomena. Comparative analysis based on Root Mean Squared Error (RMSE) metrics underscores their superiority in real-world target tracking challenges, highlighting the potential of these advanced particle filtering techniques in enhancing precision and efficiency in target tracking. Particularly, PF-ICARD achieves the highest accuracy, reducing RMSE by up to 18.64% compared to PF-CGSA and PF-IGWO, making it ideal for high-precision applications. PF-CGSA excels in specific scenarios, notably improving X-coordinate tracking in the ACT2 model by up to 11.39%. Meanwhile, PF-IGWO offers a balanced solution with moderate accuracy and enhanced computational efficiency, suitable for real-time applications. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.

关键词： Aircraft detection

来源：评论

学校读者我要写书评

暂无评论

Exploring Integration of Multimodal Deep Learning Approaches for Enhanced Alzheimer's Disease Diagnosis: A Review of Recent Literature

引用

SN computer science 2024年第7期5卷 852页

作者： Deshpande, Sonali Kulkarni, Nilima Computer Science & amp Engineering Department MIT SoC MIT Art Design and Technology University Pune India

Alzheimer's disease (AD), is the most common form of dementia that affects the nervous system. In the past few years, non-invasive early AD diagnosis has become more popular as a way to improve patient care and treatment results. Imaging methods, electroencephalogram (EEG) tests, and sound evaluations are some of the new ways that researchers have looked into. This review covers 60 papers published from 2020. They are compared in terms of how they use basic deep learning models such as CNN, LSTM, Alex Net, Inception Net, VGG19, and ResNet to identify AD. But not many studies use more than one method together, like image and EEG, EEG and sounds, or images and sounds. The information from the Scopus database makes it easy to look at the newest information and work. This means that using more than one method to find AD isn't getting as much attention. Our review says that combining the best parts of each method in a mixed way could make Alzheimer's research much more useful and lead to better ways to diagnose. The paper talks about problems and opportunities in the field right now as well as possible study topics and issues for the future. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Alzheimer's disease (AD) Deep learning Electroencephalography Machine learning Magnetic resonance imaging Neurological disorders Sound

来源：评论

学校读者我要写书评

暂无评论

Autonomous solar car 4th

Autonomous solar car

引用

4th International Conference on Computational Methods in science and Technology, ICCMST 2024

作者： Kulkarni, Nilima Sayyed, Arshadali Soni, Nikhil Pal, Komal Singh, Prashant Computer Science & Engineering Department MIT SoC MIT Art Design and Technology University Pune India

ISBN: (纸本)9781032911571

Throughout the years, automobile companies achieved outstanding progress in manufacturing safe, reliable, and affordable vehicles. Many companies and manufacturers are developing autonomous cars for the future years. However, to achieve this goal they are facing too many challenges in technical and non-technical fields like software complexity, mapping, real-time data analysis, testing, and verification. AI Software installation in cars is costly, and also it affects fuel consumption. So, proposed an autonomous solar car that will run on solar energy, which is available free of cost. NumPy, OpenCV computer vision library, AI, ML, and Python were used for the implementation of the work. The edge detection algorithm takes the input from a camera. LiDAR, WIFI/Bluetooth models, and Ultrasonic sensors have been used. Inculcated OpenCV which has an advanced built-in function that has add-on provisions in extension to the blurred image, a threshold that can go low and high and determines whether there is an edge in the path or not. The input collected from the sensors was analyzed to give directions to the car on where to get a turn. In summary, the proposed autonomous solar car model is inexpensive to build and performs well. © 2025 the Author(s).

关键词： Ultrasonic sensors

来源：评论

学校读者我要写书评

暂无评论

Fast sampling from constrained spaces using the Metropolis-adjusted Mirror Langevin algorithm 37

Fast sampling from constrained spaces using the Metropolis-a...

引用

37th Annual Conference on Learning Theory, COLT 2024

作者： Srinivasan, Vishwak Wibisono, Andre Wilson, Ashia Department of Electrical Engineering and Computer Science MIT United States Department of Computer Science Yale University United States

We propose a new method called the Metropolis-adjusted Mirror Langevin algorithm for approximate sampling from distributions whose support is a compact and convex set. This algorithm adds an accept-reject filter to the Markov chain induced by a single step of the Mirror Langevin algorithm (Zhang et al., 2020), which is a basic discretisation of the Mirror Langevin dynamics. Due to the inclusion of this filter, our method is unbiased relative to the target, while known discretisations of the Mirror Langevin dynamics including the Mirror Langevin algorithm have an asymptotic bias. For this algorithm, we also give upper bounds for the number of iterations taken to mix to a constrained distribution whose potential is relatively smooth, convex, and Lipschitz continuous with respect to a self-concordant mirror function. As a consequence of the reversibility of the Markov chain induced by the inclusion of the Metropolis-Hastings filter, we obtain an exponentially better dependence on the error tolerance for approximate constrained sampling. © 2024 V. Srinivasan, A. Wibisono & A. Wilson.

关键词： Markov chains

来源：评论

学校读者我要写书评

暂无评论

CBBM-WARM: A Workload-Aware Meta-Heuristic for Resource Management in Cloud Computing

引用

China Communications 2025年第06期 255-275页

作者： K Nivitha P Pabitha R Praveen School of Computer Science and Engineering Vellore Institute of Technology Department of Computer Technology Anna UniversityMIT Campus Department of Computer Science and Engineering National Institute of Technology Puducherry

The rapid advent in artificial intelligence and big data has revolutionized the dynamic requirement in the demands of the computing resource for executing specific tasks in the cloud environment. The process of achieving autonomic resource management is identified to be a herculean task due to its huge distributed and heterogeneous environment. Moreover, the cloud network needs to provide autonomic resource management and deliver potential services to the clients by complying with the requirements of Quality-of-Service(QoS) without impacting the Service Level Agreements(SLAs). However, the existing autonomic cloud resource managing frameworks are not capable in handling the resources of the cloud with its dynamic requirements. In this paper, Coot Bird Behavior Model-based Workload Aware Autonomic Resource Management Scheme(CBBM-WARMS) is proposed for handling the dynamic requirements of cloud resources through the estimation of workload that need to be policed by the cloud *** CBBM-WARMS initially adopted the algorithm of adaptive density peak clustering for workloads clustering of the cloud. Then, it utilized the fuzzy logic during the process of workload scheduling for achieving the determining the availability of cloud resources. It further used CBBM for potential Virtual Machine(VM) deployment that attributes towards the provision of optimal resources. It is proposed with the capability of achieving optimal Qo S with minimized time,energy consumption, SLA cost and SLA *** experimental validation of the proposed CBBMWARMS confirms minimized SLA cost of 19.21% and reduced SLA violation rate of 18.74%, better than the compared autonomic cloud resource managing frameworks.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Adapting Deep Learning Transformer Model for Depression-Specific Named Entity Recognition: A Transfer Learning Approach 3

Adapting Deep Learning Transformer Model for Depression-Spec...

引用

3rd International Conference on Automation, Computing and Renewable Systems, ICACRS 2024

作者： Pawar, Dipti Phansalkar, Shraddha Department of Computer Science Engineering MIT School of Computing MIT Art DesignTechnology University Pune India

ISBN: (纸本)9798331532420

Named Entity Recognition (NER) is essential in the biomedical domain, particularly in mental health studies focused on disorders like depression. It helps extract structured information from unstructured text, enabling accurate recognition of key entities such as symptoms, disorder, antidepressants, brands, and therapies. This step is crucial for recommending diagnoses, personalizing treatments, and advancing research. Traditional methods, like rule-based or dictionary-driven approaches, often fall short in flexibility and scalability, as they rely heavily on predefined lexicons and corpora. The proposed work presents an advanced deep learning model aimed at improving NER performance on a customized dataset. Our model focuses on identifying critical entities disorders, brands, antidepressants, and therapies by addressing the limitations of conventional techniques. It uses a hybrid approach that combines a BERT-based model, fine-tuned through transfer learning, with a knowledge-based dictionary (MedDRA). This integration allows for enhanced semantic enrichment, improving the identification and categorization of biomedical entities. By leveraging BERT’s contextual capabilities alongside the structured knowledge from MedDRA, the model delivers better accuracy and classification in mental health data. Preliminary results show high confidence levels, with substantial improvements in key metrics like precision, recall, and F1 score, further advancing NER's role in mental health research. © 2024 IEEE.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Advancing Alzheimer's Diagnosis: Performance Analysis of Deep Learning Ensembles in Medical Imaging 2

Advancing Alzheimer's Diagnosis: Performance Analysis of Dee...

引用

2nd DMIHER International Conference on Artificial Intelligence in Healthcare, Education and Industry, IDICAIEI 2024

作者： Deshpande, Sonali Kulkarni, Nilima Mit Art Design and Technology University Mit SoC Computer Science & Engineering Department Pune India

ISBN: (纸本)9798331528713

Neuroimaging and deep learning have become the focus of much research for diagnosing Alzheimer's disease (AD) in recent years. Nevertheless, the limited availability of neuroimaging training data has resulted in significant overfitting issues for several deep learning models. However, the dataset was very imbalanced. To address this issue, we employed several augmentation techniques such as flipping, rotating, and zooming to balance the dataset. The ADNI dataset consisted of four classes: mild demented, moderate demented, very mild demented, and non-demented. In this work, we have proposed an ensemble of EfficientNetB2, DenseNet121, VGG16, and Xception for the diagnosis of Alzheimer at early stage. Experiment results shows that EfficientNetB2 ensemble with Xception and VGG16 performed extremely well with 99.1% and 99.4% respectively. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Profanity Detection from Audio Recordings Using Natural Language Processing Techniques

Profanity Detection from Audio Recordings Using Natural Lang...

引用

International Conference on Recent Developments in Cyber Security, ReDCySec 2023

作者： Patil, Swapnil Chandre, Pankaj R. Vanarote, Viresh Pathan, Shafi Nimbalkar, Madhukar Shinde, Gitanjali Department of Computer Science and Engineering MIT School of Computing MIT Art Design and Technology University Loni Kalbhor India Department of Computer Science and Engineering Vishwakarma Institute of Information Technology Kondhwa India

ISBN: (纸本)9789819998104

Profanity detection has become increasingly important in various industries, including media and online content moderation. In this paper, we propose a novel approach to identifying profane words from audio using natural language processing (NLP) techniques. We first convert the audio files into textual transcripts using automatic speech recognition (ASR) tools. We then apply pre-processing techniques such as tokenization, stemming, and stop-word removal to the transcripts. Finally, we train various machine learning models, such as Naive Bayes, Support Vector Machines (SVMs), and Random Forests, to classify the transcripts as either containing profane language or not. The proposed approach can be integrated into various applications, such as content moderation tools, online chat systems, and automatic transcription services. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Audio recordings

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：