Article
Environmental Sciences
Feng Gao, Wei Zhang, Andrea A. Baccarelli, Yike Shen
Summary: The in-silico prediction of chemical ecotoxicity (HC50) plays a crucial role in enhancing toxicological assessment of manufactured chemicals. A novel autoencoder model was developed to learn latent chemical representations, achieving state-of-the-art prediction performance for HC50.
ENVIRONMENT INTERNATIONAL
(2022)
Article
Environmental Sciences
Elham Rafiei Sardooi, Hamid Reza Pourghasemi, Ali Azareh, Farshad Soleimani Sardoo, John J. Clague
Summary: This study explored the prediction of ground subsidence using various statistical and machine learning models in the Rafsanjan Plain in Iran, highlighting the importance of factors such as NDVI, groundwater drawdown, land use, and lithology. The SVM model showed the highest prediction accuracy among the tested models.
GEOCARTO INTERNATIONAL
(2022)
Article
Chemistry, Multidisciplinary
Vishwesh Venkatraman
Summary: This article examines the efficacy of fingerprint-based machine learning models for a large number of ADMET-related properties. For most properties, fingerprint-based random forest models show comparable or better performance than traditional 2D/3D molecular descriptors.
JOURNAL OF CHEMINFORMATICS
(2021)
Article
Engineering, Civil
Zhong-kai Feng, Wen-jing Niu, Zheng-yang Tang, Yang Xu, Hai-rong Zhang
Summary: A novel evolutionary artificial intelligence model is developed for multiple scales nonstationary hydrological time series prediction, utilizing the cooperation search algorithm to optimize the ELM model's input-hidden weights and biases. Experimental results show that the proposed method outperforms the traditional ELM method in terms of performance evaluation indexes, particularly with significant improvements in both RMSE and MAPE during testing phase, supporting decision-making in water resource system.
JOURNAL OF HYDROLOGY
(2021)
Article
Computer Science, Artificial Intelligence
Puck van Gerwen, Alberto Fabrizio, Matthew D. Wodrich, Clemence Corminboeuf
Summary: Physics-based representations (QML) can reliably and efficiently infer molecular properties, and we have extended its capabilities to predict reaction properties. By defining reaction representations and utilizing existing molecular representations, we can take multiple molecules participating in a reaction as input. We also introduce a new dataset for benchmarking and evaluating the performance of reaction representations.
MACHINE LEARNING-SCIENCE AND TECHNOLOGY
(2022)
Article
Automation & Control Systems
Indrajeet Kumar, Bineet Kumar Tripathi, Anugrah Singh
Summary: Petroleum production forecasting involves predicting fluid production from wells using historical data. Traditional methods and conventional machine learning techniques are time-consuming and have limited forecasting power. In this study, we developed time-series forecast models based on an attention-based long short-term memory network, which outperforms other models for petroleum production forecasting.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Operations Research & Management Science
Md. Iftekharul Alam Efat, Petr Hajek, Mohammad Zoynul Abedin, Rahat Uddin Azad, Md. Al Jaber, Shuvra Aditya, Mohammad Kabir Hassan
Summary: Existing sales forecasting models are not comprehensive and flexible enough to consider dynamic changes and nonlinearities in sales time-series. This study proposes a hybrid model that combines adaptive trend estimated series (ATES) with a deep neural network model to capture different big data characteristics in sales forecasting data. The proposed hybrid model outperforms existing models for forecasting horizons ranging from one to 12 months.
ANNALS OF OPERATIONS RESEARCH
(2022)
Article
Engineering, Multidisciplinary
Simone Massulini Acosta, Anderson Levati Amoroso, Angelo Marcio Oliveira Sant Anna, Osiris Canciglieri Junior
Summary: Relevance vector machines, a Bayesian sparse kernel method, coupled with a self-adaptive differential evolution algorithm, show superior performance in predicting phosphorus concentration levels in the steelmaking process. The study indicates that RVM models are an adequate tool for such predictions.
APPLIED MATHEMATICAL MODELLING
(2021)
Article
Biochemistry & Molecular Biology
Baddipadige Raju, Himanshu Verma, Gera Narendra, Bharti Sapra, Om Silakari
Summary: The study utilized in-silico approaches to identify selective CYP1B1 inhibitors, screening for the most stable inhibitors through molecular docking analysis, which may offer a new avenue for addressing resistance in tumors.
JOURNAL OF BIOMOLECULAR STRUCTURE & DYNAMICS
(2022)
Article
Green & Sustainable Science & Technology
Arnold E. Sison, Sydney A. Etchieson, Fatih Gulec, Emmanuel I. Epelle, Jude A. Okolie
Summary: Chemical looping gasification (CLG) is an advanced thermochemical process that utilizes solid metal oxides as oxidants to transfer oxygen from an air reactor to a gasification reactor, resulting in the production of hydrogen gas with a smaller carbon footprint compared to conventional gasification. However, CLG still faces challenges such as high capital cost, durability of oxygen carriers, complex reaction mechanism, and scalability issues. This study proposes a novel approach combining process simulation, experimental studies, and machine learning analysis to predict hydrogen and char yield during CLG, with gradient boost regression (GBR) outperforming other models.
JOURNAL OF CLEANER PRODUCTION
(2023)
Article
Computer Science, Artificial Intelligence
Yupeng Hu, Peng Zhan, Yang Xu, Jia Zhao, Yujun Li, Xueqing Li
Summary: Recent years have seen significant growth in time series data due to the popularity of sensing devices and IoT techniques, making time series classification one of the most challenging studies in data mining. Empirical evidence suggests that learning-based time series classification methods have advantages in accuracy, efficiency, and interpretability compared to traditional methods. However, the high time complexity of feature processing has limited the performance of these methods. This paper introduces an efficient shapelet transformation method and an enhanced recurrent neural network model for deep representation learning to improve the overall efficiency and accuracy of time series classification.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Geosciences, Multidisciplinary
Pin Zhang, Zhen-Yu Yin, Yin-Fu Jin, Tommy H. T. Chan, Fu-Ping Gao
Summary: This study proposes a novel modeling approach using machine learning techniques to predict the compression index C c in geotechnical design, showing that machine learning models outperform traditional empirical prediction formulations. Among the tested machine learning algorithms, random forest and back-propagation neural network models are recommended for predicting C c under different conditions.
GEOSCIENCE FRONTIERS
(2021)
Article
Automation & Control Systems
Di Wu, Qinghua Guan, Zhe Fan, Hanhui Deng, Tao Wu
Summary: With the development of AI and hardware computing power, deep learning models are widely used in IoT for analyzing spatiotemporal data from wireless sensors. Manual hyperparameter optimization is costly and may result in unreasonable settings and poor performance. This article proposes an automated HPO method based on parallel genetic algorithm for LSTM models. The method is divided into stages such as population initialization, fitness function, selection, crossover operators, mutation operators, subgroup exchange, and end of evolution. Experimental results show that the proposed method outperforms other mainstream HPO methods in terms of time costs and prediction results.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Chemistry, Multidisciplinary
Shinya Inazumi, Sudip Shakya, Chifong Chio, Hideki Kobayashi, Supakij Nontananandh
Summary: In the field of geotechnical engineering, the use of chemical injection as a ground improvement method for mitigating liquefaction and land subsidence problems is of great concern. This study evaluates the compressive strength of improved grounds using granulated blast furnace slag and develops a time-series prediction model to assess long-term sustainability. Three models, ARIMA, SSR, and MLP, are compared and the MLP model proves to be the most reliable for long-term prediction when sufficient input information is available. However, the SSR model performs better overall with scarce input data, while the ARIMA model generates higher errors.
APPLIED SCIENCES-BASEL
(2023)
Article
Polymer Science
Xiaoyu Huang, Shuai Wang, Tong Lu, Houmin Li, Keyang Wu, Weichao Deng
Summary: The addition of rubber to concrete improves resistance to chloride ion attacks, and determining the chloride permeability coefficient (D-CI) of rubber concrete (RC) is significant for coastal areas. This paper presents a mixed whale optimization algorithm (MWOA) to optimize machine learning models, which enhances the prediction accuracy of D-CI of RC.
Article
Chemistry, Medicinal
Christin Rakers, Rifat Ara Najnin, Ahsan Habib Polash, Shunichi Takeda, J. B. Brown
Article
Chemistry, Medicinal
J. B. Brown
MOLECULAR INFORMATICS
(2018)
Article
Biochemistry & Molecular Biology
Ahsan Habib Polash, Takumi Nakano, Shunichi Takeda, J. B. Brown
Article
Chemistry, Physical
Stephanie Maria Linker, Aniket Magarkar, Juergen Koefinger, Gerhard Hummer, Daniel Seeliger
JOURNAL OF CHEMICAL THEORY AND COMPUTATION
(2019)
Article
Chemistry, Medicinal
Ferruccio Palazzesi, Markus R. Hermann, Marc A. Grundl, Alexander Pautsch, Daniel Seeliger, Christofer S. Tautermann, Alexander Weber
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2020)
Article
Chemistry, Medicinal
Sebastian W. Draxler, Margit Bauer, Christian Eickmeier, Simon Nadal, Herbert Nar, Daniel Rangel Rojas, Daniel Seeliger, Markus Zeeb, Dennis Fiegen
JOURNAL OF MEDICINAL CHEMISTRY
(2020)
Article
Biology
Yuki Makino, Yuki Kamiyama, J. B. Brown, Toshiya Tanaka, Ryusuke Murakami, Yuki Teramoto, Takayuki Goto, Shusuke Akamatsu, Naoki Terada, Takahiro Inoue, Tatsuhiko Kodama, Osamu Ogawa, Takashi Kobayashi
Summary: The study reveals that OPRK1 plays an important role in the progression of castration resistance in prostate cancer. Loss of OPRK1 function can delay the acquisition of castration resistance and inhibit the growth of castration-resistant prostate cancer.
COMMUNICATIONS BIOLOGY
(2022)
Correction
Biology
Yuki Makino, Yuki Kamiyama, J. B. Brown, Toshiya Tanaka, Ryusuke Murakami, Yuki Teramoto, Takayuki Goto, Shusuke Akamatsu, Naoki Terada, Takahiro Inoue, Tatsuhiko Kodama, Osamu Ogawa, Takashi Kobayashi
COMMUNICATIONS BIOLOGY
(2022)
Article
Oncology
Shiro Takamatsu, J. B. Brown, Ken Yamaguchi, Junzo Hamanishi, Koji Yamanoi, Hisamitsu Takaya, Tomoko Kaneyasu, Seiichi Mori, Masaki Mandai, Noriomi Matsumura
Summary: This study comprehensively analyzed the association between HRR pathway gene alterations and genomic scar scores, and found that biallelic alterations in HRR genes other than BRCA1/2 were also associated with elevated genomic scar scores. The combination of these indices can be used to identify HRD cases and provide better prognosis when treated with DNA-damaging agents.
JCO PRECISION ONCOLOGY
(2022)
Article
Oncology
Shiro Takamatsu, J. B. Brown, Ken Yamaguchi, Junzo Hamanishi, Koji Yamanoi, Hisamitsu Takaya, Tomoko Kaneyasu, Seiichi Mori, Masaki Mandai, Noriomi Matsumura
Summary: Biallelic alterations in HRR genes other than BRCA1/2 are associated with elevated genomic scar scores, and this association varies significantly by sex and the presence of somatic TP53 mutations. Tumors with HRD features in gene expression analysis due to a combination of indices show significantly higher sensitivity to DNA-damaging agents, both in clinical samples and cell lines. This study supports the utility of HRD analysis in all cancer types and enhances chemotherapy decision making and efficacy in clinical settings, marking a significant advancement in precision oncology.
JCO PRECISION ONCOLOGY
(2021)
Article
Biochemistry & Molecular Biology
Takumi Nakano, Shunichi Takeda, J. B. Brown
RSC MEDICINAL CHEMISTRY
(2020)
Article
Chemistry, Multidisciplinary
Vytautas Gapsys, Laura Perez-Benito, Matteo Aldeghi, Daniel Seeliger, Herman Van Vlijmen, Gary Tresadern, Bert L. de Groot