检索结果-内蒙古大学图书馆

Estimating the robustness of classification models by the structure of the learned feature-space

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Ho, Kalun Pfreundt, Franz-Josef Keuper, Janis Keuper, Margret CC-HPC Fraunhofer ITWM Fraunhofer-Platz 1 Kaiserslautern67663 Germany Institute for Machine Learning and Analytics Offenburg University Germany Data and Web Science Group University of Mannheim Germany

Over the last decade, the development of deep image classification networks has mostly been driven by the search for the best performance in terms of classification accuracy on standardized benchmarks like ImageNet. More recently, this focus has been expanded by the notion of model robustness, i.e. the generalization abilities of models towards previously unseen changes in the data distribution. While new benchmarks, like ImageNet-C, have been introduced to measure robustness properties, we argue that fixed testsets are only able to capture a small portion of possible data variations and are thus limited and prone to generate new overfitted solutions. To overcome these drawbacks, we suggest to estimate the robustness of a model directly from the structure of its learned feature-space. We introduce robustness indicators which are obtained via unsupervised clustering of latent representations from a trained classifier and show very high correlations to the model performance on corrupted test data. © 2021, CC BY.

关键词： Classification (of information)

Ultrasound-based Kidney Disease Detection: Resampling Cubature Kalman Filter and Physics-Guided VGG Cross-Contextual Network 8

学校读者我要写书评

暂无评论

Ultrasound-based Kidney Disease Detection: Resampling Cubatu...

8th International Conference on Inventive Computation Technologies, ICICT 2025

作者： Padmanayana Kurumeti, N.S.L. Kumar Veena, S. Raja, Ch. Upadhye, Vijay Jagdish Maranan, Ramya Department of Artificial Intelligence and Data Science Engineering AJ Institute of Engineering and Technology Karnataka Mangaluru 575006 India Department of Artificial Intelligence and Machine Learning Aditya University Andhra Pradesh Surampalem 533437 India Department of Electronics and Communication Engineering SJC Institute of Technology Karnataka Chikkaballapur 562101 India Department of Electronics and Communication Engineering Mahatma Gandhi Institute of Technology Telangana Hyderabad 500075 India Department of Microbiology Research and Development Cell (RDC) Dept of Microbiology Parul Institute of Applied Sciences (PIAS) Parul University Post Limda Gujarat Waghodia 391760 India Department of Research and Innovation Saveetha School of Engineering SIMATS Tamil Nadu Chennai 602105 India

ISBN: (纸本)9798331512248

Limitations of ultrasound-based identification of kidney disease include poor image resolution, operator reliance, and lack of detection of small tissue anomalies. Diagnostic accuracy may be affected by artefacts or noise, or changes in patient anatomy. It may also have difficulty detecting microscopic structural changes or renal disease in its incipiency. The limitations are overcome by the analysis of the Resampling Cubature Kalman Filter and Physics-Guided VGG Cross-Contextual Networks (RCKF-PG-VGG-CCNet) for kidney disease detection from ultrasound images based on the Kidney Ultrasound Image dataset. The initial use of the Resampling Cubature Kalman Filter and the first use of Physics-Guided VGG Cross-Contextual Networks is to enhance image quality. Through identifying regions through deviations, the Scale-Aware Meet Transformer (SAMT) performs segmentation. While the Synergistic Fibroblast Optimization (SFO) enhances classification performance through optimization of the search for diseased regions, the Physics-Guided VGG Cross-Contextual Networks (PG-VGG-CCNet) classify kidney disease and identify features. Experimental output from a Python-based experiment demonstrates that RCKF-PG-VGG-CCNet is more efficient than other methods with 99.9% efficiency and 99.8% sensitivity and may enhance the accuracy of kidney disease detection. © 2025 IEEE.

关键词： Kidney disease Kidney Ultrasound Image dataset Resampling Cubature Kalman Filter Scale-Aware Meet Transformer Synergistic Fibroblast Optimization

Active learning meets fractal decision boundaries: a cautionary tale from the Sitnikov three-body problem

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Payot, Nicolas Pasquato, Mario Trani, Alessandro Alberto Hezaveh, Yashar Perreault-Levasseur, Laurence Département de Physique Université de Montréal Mila Quebec Artificial Intelligence Institute Ciela - Montreal Institute for Astrophysical Data Analysis and Machine Learning Canada Dipartimento di Fisica e Astronomia Università di Padova Vicolo dell’Osservatorio 5 PadovaI-35122 Italy Niels Bohr Institute Copenhagen Denmark Research Center for the Early Universe School of Science The University of Tokyo Tokyo Japan Okinawa Institute of Science and Technology Okinawa Japan

Chaotic systems such as the gravitational N-body problem are ubiquitous in astronomy. machine learning (ML) is increasingly deployed to predict the evolution of such systems, e.g. with the goal of speeding up simulations. Strategies such as active learning (AL) are a natural choice to optimize ML training. Here we showcase an AL failure when predicting the stability of the Sitnikov three-body problem, the simplest case of N-body problem displaying chaotic behavior. We link this failure to the fractal nature of our classification problem’s decision boundary. This is a potential pitfall in optimizing large sets of N-body simulations via AL in the context of star cluster physics, galactic dynamics, or cosmology. © 2023, CC BY.

关键词： Fractals

Uncertainty Quantification of the Virial Black Hole Mass with Conformal Prediction

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Yong, Suk Yee Ong, Cheng Soon Australia Astrophysics and Space Technologies Research Centre Macquarie University SydneyNSW2109 Australia School of Mathematical and Physical Sciences Macquarie University SydneyNSW2109 Australia Faculty of Science and Engineering Macquarie University SydneyNSW2109 Australia Machine Learning and Artificial Intelligence Future Science Platform CSIRO Australia Data61 CSIRO CanberraACT2601 Australia School of Computing The Australian National University CanberraACT2601 Australia

Precise measurements of the black hole mass are essential to gain insight on the black hole and host galaxy co-evolution. A direct measure of the black hole mass is often restricted to nearest galaxies and instead, an indirect method using the single-epoch virial black hole mass estimation is used for objects at high redshifts. However, this method is subjected to biases and uncertainties as it is reliant on the scaling relation from a small sample of local active galactic nuclei. In this study, we propose the application of conformalised quantile regression (CQR) to quantify the uncertainties of the black hole predictions in a machine learning setting. We compare CQR with various prediction interval techniques and demonstrated that CQR can provide a more useful prediction interval indicator. In contrast to baseline approaches for prediction interval estimation, we show that the CQR method provides prediction intervals that adjust to the black hole mass and its related properties. That is it yields a tighter constraint on the prediction interval (hence more certain) for a larger black hole mass, and accordingly, bright and broad spectral line width source. Using a combination of neural network model and CQR framework, the recovered virial black hole mass predictions and uncertainties are comparable to those measured from the Sloan Digital Sky Survey. The code is publicly available at https://***/yongsukyee/uncertain_blackholemass. Copyright © 2023, The Authors. All rights reserved.

关键词： Forecasting

Automated question-answer medical model based on deep learning technology 20

学校读者我要写书评

暂无评论

Automated question-answer medical model based on deep learni...

6th International Conference on Engineering and MIS, ICEMIS 2020

作者： Abdallah, Abdelrahman Kasem, Mahmoud Hamada, Mohamed A. Sdeek, Shaymaa MSc Machine Learning and Data Science Satbayev University Almaty Kazakhstan Information Technology Assiut University Assiut Egypt IS Department International IT University Almaty Kazakhstan Dept. Information System Assiut University Qena Egypt

ISBN: (纸本)9781450377362

Artificial intelligence can now provide more solutions for different problems, especially in the medical field. One of those problems is the lack of answers to any given medical/health-related question. The Internet is full of forums that allow people to ask some specific questions and get great answers for them. Nevertheless, browsing these questions to locate a similar case to your own question, also finding a satisfying accurate answer is difficult and timeconsuming task. This research will introduce a solution to these problems by automating the process of generating qualified answers to these questions and creating a kind of digital doctor. Furthermore, this research will train an end-to-end model using the framework of RNN and the encoder decoder to generate sensible and useful answers to a small set of medical/health issues. The proposed model was trained and evaluated using data from various online services, such as WebMD, HealthTap, eHealthForums, and iCliniq. © 2020 ACM.

关键词： Deep learning

Myocarditis Diagnosis: A Method using Mutual learning-Based ABC and Reinforcement learning

学校读者我要写书评

暂无评论

Myocarditis Diagnosis: A Method using Mutual Learning-Based ...

International Symposium on Computational Intelligence and Informatics

作者： Saba Danaei Arsam Bostani Seyed Vahid Moravvej Fardin Mohammadi Roohallah Alizadehsani Afshin Shoeibi Hamid Alinejad-Rokny Saeid Nahavandi Adiban Institute of Higher Education Semnan Iran Department of mechanical engineering of biosystems Urmia university Department of exercise physiology & health science University of tehran Internship in UNSW BioMedical Machine Learning Lab Sydney NSW Australia Institute for Intelligent Systems Research and Innovation (IISRI) Deakin University Waurn Ponds Victoria Australia UNSW Data Science Hub The University of New South Wales (UNSW Sydney) Sydney New South Wales Australia BioMedical Machine Learning Lab The Graduate School of Biomedical Engineering UNSW Sydney Sydney NSW Australia

ISBN: (纸本)9798350398830

Myocarditis occurs when the heart muscle becomes inflamed and inflammation occurs when your body’s immune system responds to infections. It can be diagnosed using cardiac magnetic resonance image (MRI), a non-invasive imaging technique with the possibility of operator bias. This paper proposes a hybrid method of deep reinforcement learning-based algorithms and meta-heuristics algorithms. A mutual learning-based artificial bee colony (ML-ABC) is employed for initial weight, which adjusts the candidate food source generated with the higher fitness between two individuals determined by a mutual learning factor. Moreover, a sequential decision-making process investigates the imbalanced classification issue, in which a convolutional neural network (CNN) is used as the foundation for policy architecture. At first, initial weights are produced using the ML-ABC algorithm. After that, the agent receives a sample at each phase and classifies it, obtaining environmental rewards. The minority class receives more rewards than the majority class. Eventually, the agent discovers an ideal strategy with the aid of a specific reward function and a beneficial learning environment. We evaluate our proposed approach on the Z-Alizadeh Sani myocarditis dataset based on standard criteria and demonstrate that the proposed method gives superior myocarditis diagnosis performance.

关键词： Mechatronics Computational modeling Metaheuristics Reinforcement learning Myocardium Markov processes Classification algorithms

Reinforcement learning Algorithm for Mixed Mean Field Control Games

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Angiuli, Andrea Detering, Nils Fouque, Jean-Pierre Laurière, Mathieu Lin, Jimin Prime Machine Learning Team Amazon. 320 Westlake Ave N SEA83 SeattleWA98109 United States Department of Statistics and Applied Probability South Hall University of California Santa BarbaraCA93106 United States Mathematics and Data Science NYU Shanghai China

We present a new combined mean field control game (MFCG) problem which can be interpreted as a competitive game between collaborating groups and its solution as a Nash equilibrium between groups. Players coordinate their strategies within each group. An example is a modification of the classical trader's problem. Groups of traders maximize their wealth. They face cost for their transactions, for their own terminal positions, and for the average holding within their group. The asset price is impacted by the trades of all agents. We propose a three-timescale reinforcement learning algorithm to approximate the solution of such MFCG problems. We test the algorithm on benchmark linear-quadratic specifications for which we provide analytic *** Codes 91A16, 68Q32 © 2022, CC BY.

关键词： Commerce

The Alchemical Integral Transform revisited

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Krug, Simon León von Lilienfeld, O. Anatole Machine Learning Group Technische Universität Berlin Berlin10587 Germany Berlin Institute for the Foundations of Learning and Data Berlin10587 Germany Chemical Physics Theory Group Department of Chemistry University of Toronto St. George Campus TorontoON Canada Department of Materials Science and Engineering University of Toronto St. George Campus TorontoON Canada Vector Institute for Artificial Intelligence TorontoON Canada Department of Physics University of Toronto St. George Campus TorontoON Canada Acceleration Consortium University of Toronto TorontoON Canada

We recently introduced the Alchemical Integral Transform (AIT) enabling the prediction of energy differences, and guessed an Ansatz to parametrize space r in some alchemical change λ. Here, we present a rigorous derivation of AIT’s kernel K and discuss the parametrization r(λ) in n dimensions, i.e. necessary conditions, mathematical freedoms and additional constraints when obtaining it. Analytical expressions for changes in energy spectra and densities are given for a number of systems. Examples include homogeneous potentials like the quantum harmonic oscillator, Hydrogen-like atom, and Dirac well, both for one- and multiparticle cases, and a multiparticle system beyond coordinate scaling for harmonic potentials. © 2023, CC BY.

关键词： Choquet integral

Weakly-Supervised Depression Detection in Speech Through Self-learning Based Label Correction

学校读者我要写书评

暂无评论

IEEE Transactions on Audio, Speech and Language Processing

IEEE Transactions on Audio, Speech and Language Processing 2025年 33卷 748-758页

作者： Yanfei Sun Yuanyuan Zhou Xinzhou Xu Jin Qi Feiyi Xu Zhao Ren Björn W. Schuller School of Internet of Things Nanjing University of Posts and Telecommunications Nanjing China Wuxi University Wuxi China School of Computer Science Nanjing University of Posts and Telecommunications Nanjing China Signal Processing and Speech Communication Laboratory Graz University of Technology Graz Austria Key Laboratory of Modern Acoustics MOE Nanjing University Nanjing China Cognitive Systems Lab University of Bremen Bremen Germany Chair of Health Informatics Technische Universität München (TUM) München Germany Munich Data Science Institute Munich Germany Munich Center for Machine Learning Munich Germany GLAM – the Group on Language Audio & Music Imperial College London London U.K.

Automated Depression Detection (ADD) in speech aims to automatically estimate one's depressive attributes through artificial intelligence tools towards spoken signals. Nevertheless, existing speech-based ADD works fail to sufficiently consider weakly-supervised cases with inaccurate labels, which may typically appear in intelligent mental health. In this regard, we propose the Self-learning-based Label Correction (SLLC) approach for weakly-supervised depression detection in speech. The proposed approach employs a self-learning manner connecting a label correction module and a depression detection module. Within the approach, the label correction module fuses likelihood-ratio-based and prototype-based label correction strategies in order to effectively correct the inaccurate labels, while the depression detection module aims at detecting depressed samples through a 1D convolutional recurrent neural network with multiple types of losses. The experimental results on two depression detection corpora show that our proposed SLLC approach performs better compared with existing state-of-the-art speech-based depression detection approaches, in the case of weak supervision with inaccurate labels for depression detection in speech.

关键词： Depression Feature extraction Training Speech processing Prototypes Noise measurement Telecommunications Sun Internet of Things Weak supervision