Article
Chemistry, Physical
Dennis P. P. Trujillo, Ashok Gurung, Jiacheng Yu, Sanjeev K. K. Nayak, S. Pamir Alpay, Pierre-Eymeric Janolin
Summary: A study combining data mining and first-principles computations has discovered a group of iodides, bromides, and chlorides with high electrostrictive coefficients and effective piezoelectric voltage coefficients.
NPJ COMPUTATIONAL MATERIALS
(2022)
Article
Engineering, Electrical & Electronic
Jingjing Wen, Houpu Yao, Ze Ji, Bin Wu, Feng Xu
Summary: This study introduces a machine learning system for self-validation of high-g accelerometers to ensure the reliability of pyroshock tests for space electronics. By combining ensemble learning model and deep neural network, the system successfully identifies the health conditions and fault types of damaged accelerometers, as well as recovers corrupted shock signals.
SENSORS AND ACTUATORS A-PHYSICAL
(2021)
Article
Physics, Particles & Fields
Andrew Chisholm, Thomas Neep, Konstantinos Nikolopoulos, Rhys Owen, Elliot Reynolds, Julia Silva
Summary: Background modelling is a major challenge in particle physics data analysis. This paper proposes a widely applicable approach for non-parametric data-driven background modelling, which addresses the limitations of simulated events and parametric models. The approach relies on a relaxed event selection to estimate conditional probability density functions, and two different techniques are discussed for its realization.
JOURNAL OF HIGH ENERGY PHYSICS
(2022)
Article
Computer Science, Artificial Intelligence
Achuta Kadambi, Celso de Melo, Cho-Jui Hsieh, Mani Srivastava, Stefano Soatto
Summary: Computer vision techniques, often data-driven, can benefit from including physical models as constraints in the pipeline. This Perspective provides an overview of specific approaches for integrating physics into artificial intelligence pipelines, referred to as physics-based machine learning.
NATURE MACHINE INTELLIGENCE
(2023)
Review
Energy & Fuels
Wendi Guo, Zhongchao Sun, Soren Byg Vilsen, Jinhao Meng, Daniel Ioan Stroe
Summary: This paper presents a review of various approaches to lithium-ion battery lifetime prediction models, combining physics-based models and data-driven models. The study finds that electrochemical models and machine learning have great potential in predicting the lifespan of lithium-ion batteries, but there are also limitations and challenges in their applications.
JOURNAL OF ENERGY STORAGE
(2022)
Article
Engineering, Electrical & Electronic
Rui Guo, Tianyao Huang, Maokun Li, Haiyang Zhang, Yonina C. Eldar
Summary: Electromagnetic (EM) imaging is widely used in various fields, but it is an ill-posed inverse problem. Machine learning techniques, particularly deep learning, have shown potential in fast and accurate imaging. However, the challenge lies in constructing a training set that accurately represents practical scenarios. To overcome this, recent research has focused on physics-embedded ML methods for EM imaging, which combine the benefits of big data and the theoretical constraints of physical laws.
IEEE SIGNAL PROCESSING MAGAZINE
(2023)
Article
Mathematics, Applied
Kaikai Cao, Xiaochen Zeng
Summary: This paper proposes a data-driven wavelet estimator for deconvolution density model. Additionally, we investigate the totally adaptive estimations with moderately ill-posed noises on Besov spaces B-r,q(s)(R). The estimation for the case of 0 < s <= 1/r is considered, and the convergence rate in the region of 1 <= p <= 2sr+(2 beta+1)r/sr+2 beta+1 is improved compared to not necessarily compactly supported density estimations.
RESULTS IN MATHEMATICS
(2023)
Article
Engineering, Electrical & Electronic
Wenjun Xia, Hongming Shan, Ge Wang, Yi Zhang
Summary: Since 2016, deep learning has made remarkable progress in tomographic imaging, particularly in low-dose computed tomography (LDCT). However, the black-box nature and instabilities of LDCT denoising and end-to-end reconstruction networks hinder the application of DL methods in LDCT. A recent trend is to integrate imaging physics and models into deep networks, allowing for a combination of physics-/model-based and data-driven elements. This article provides a systematic review of physics-/model-based data-driven methods for LDCT, including loss functions, training strategies, performance evaluation, and future directions.
IEEE SIGNAL PROCESSING MAGAZINE
(2023)
Article
Engineering, Geological
Q. J. Pan, X. Z. Li, S. Y. Wang, K. K. Phoon
Summary: This study performs a statistically rigorous model comparison and a probabilistic assessment of information gain for different types of monitoring data in tunneling-induced ground deformation analysis. The results indicate that the Loganathan-Poulos model is the most suitable for predicting tunneling-induced ground deformations. The analysis also reveals that measured ground vertical deformations are more informative than measured ground horizontal deformations.
Article
Automation & Control Systems
Edward O'Dwyer, Eric C. Kerrigan, Paola Falugi, Marta Zagorowska, Nilay Shah
Summary: This article proposes a restructuring method based on segmented prediction trajectories, which can reduce the modeling burden and tracking error in control design, and provide consistent performance. Case studies show that the method can achieve good tracking performance in building energy management problems.
IEEE TRANSACTIONS ON CONTROL SYSTEMS TECHNOLOGY
(2023)
Article
Energy & Fuels
Kadir Amasyali, Nora M. El-Gohary
Summary: This study proposes a real data-driven method using machine learning and optimization models to evaluate and optimize occupant behavior for reducing energy consumption and improving comfort.
Article
Physics, Multidisciplinary
Francesco Mirani, Daniele Calzolari, Arianna Formenti, Matteo Passoni
Summary: Laser-driven photon activation analysis (PAA) is proposed as a new method for material characterization, utilizing high-energy photons for analysis. Theoretical simulations were used to identify optimal experimental conditions for laser-driven PAA, showing comparable performance with conventional accelerators under high repetition rate operation.
COMMUNICATIONS PHYSICS
(2021)
Article
Construction & Building Technology
Syed Ahsan Raza Naqvi, Koushik Kar, Sandipan Mishra
Summary: This paper investigates the temperature control in shared workspaces with different heating and cooling sources for energy saving and personalized environment. It proposes multiple time-bound control strategies for preparing the workspace before scheduled activities and a separate control strategy for enhancing occupant comfort during occupied intervals. Experimental results show that the proposed strategies significantly save energy and achieve the desired indoor temperature.
ENERGY AND BUILDINGS
(2023)
Article
Environmental Sciences
Anil Kumar, Panagiotis Kosmopoulos, Yashwant Kashyap, Rupam Gautam
Summary: In this study, the possibility of estimating GHI in parallel to PV power production in India was investigated using the RTM model called libRadtran. Satellite information on cloud and aerosol conditions, along with ground-based measurements of GHI and COT, were used as input parameters. The simulation results were compared with actual data from four solar power plants in Rajasthan, India. The study found significant attenuation due to clouds and aerosols, with a maximum energy loss observed at one location.
Article
Physiology
Amel Karoui, Mostafa Bendahmane, Nejib Zemzemi
Summary: By comparing data-driven and traditional methods for cardiac arrhythmia diagnosis, the study found that the data-driven approach outperforms the classic technique and demonstrates greater robustness against noise.
FRONTIERS IN PHYSIOLOGY
(2021)
Article
Physics, Multidisciplinary
Suyong Choi, Bogyeom Kim
Summary: Measurements of the Cabbibo-Kobayashi-Maskawa matrix elements are crucial for testing the standard model, with some discrepancies identified in current measurements of |V-cb|. A new method proposed for measuring |V-cb at the LHC shows promise in resolving these discrepancies, utilizing real W bosons from tt-bar events for analysis.
JOURNAL OF THE KOREAN PHYSICAL SOCIETY
(2021)
