Amir Mosavi | German Research Center for Artificial Intelligence (original) (raw)

Papers by Amir Mosavi

Large language models (LLMs) provide competitive advantages to various fields of research. This s... more Large language models (LLMs) provide competitive advantages to various fields of research. This survey explores the journey from using deep learning to adopting ChatGPT in materials design. It covers how methodologies have shifted, moving away from traditional deep learning models to embracing LLMs such as ChatGPT, in advancing materials design research. The focus is on methods and applications, highlighting how LLMs are shaping the landscape of materials design. The survey aims to provide insights into the essential role of ChatGPT in this domain, offering a comprehensive view of its methods and diverse applications.

Mammography is often used as the most common laboratory method for the detection of breast cancer... more Mammography is often used as the most common laboratory method for the detection of breast cancer, yet associated with the high cost and many side effects. Machine learning prediction as an alternative method has shown promising results. This paper presents a method based on a multilayer fuzzy expert system for the detection of breast cancer using an extreme learning machine (ELM) classification model integrated with radial basis function (RBF) kernel called ELM-RBF, considering the Wisconsin dataset. The performance of the proposed model is further compared with a linear-SVM model. The proposed model outperforms the linear-SVM model with RMSE, R 2 , MAPE equal to 0.1719, 0.9374 and 0.0539, respectively. Furthermore, both models are studied in terms of criteria of accuracy, precision, sensitivity, specificity, validation, true positive rate (TPR), and false-negative rate (FNR). The ELM-RBF model for these criteria presents better performance compared to the SVM model.

Deep learning (DL) is a promising technology for enhancing the development of fifth generation (5... more Deep learning (DL) is a promising technology for enhancing the development of fifth generation (5G) and sixth generation (6G) mobile networks, as it can improve their capabilities, security, and performance. However, there are still significant challenges to be addressed in the implementation of DL techniques in these networks. To address these challenges, we conducted a systematic review of the literature on DL techniques in 5G and 6G applications following the PRISMA guidelines. The review was conducted in three stages: data collection, analysis, and reporting of primary findings. After evaluating and reviewing the databases, we found that hybrid DL and ensemble techniques show promise in optimizing 5G and 6G networks, given proper implementation. Finally, we discussed the open issues and challenges in this field. This review provides important insights into the potential of DL techniques in improving 5G and 6G networks, and it highlights the need for further research to overcome the remaining challenges. The results of this primary communication will be further developed and extended into a journal article.

Diabetes mellitus (DM) is a major health concern among children with the widespread adoption of a... more Diabetes mellitus (DM) is a major health concern among children with the widespread adoption of advanced technologies. However, concerns are growing about the transparency, replicability, biasedness, and overall validity of artificial intelligence studies in medicine.

The fast advancement of technology and developers’ utilization of data centers have dramatically ... more The fast advancement of technology and developers’ utilization of data centers have dramatically increased energy usage in today’s society. Thermal control is a key issue in hyper-scale cloud data centers. Hot spots form when the temperature of the host rises, increasing cooling costs and affecting dependability. Precise estimation of host temperatures is critical for optimal resource management.Thermal changes in the data center make estimating temperature a difficult challenge. Existing temperature estimating algorithms are ineffective due to their processing complexity as well as lack of accuracy. Regarding that data-driven approaches seem promising for temperature prediction, this research offers a unique efficient temperature prediction model. The model uses a combination of convolutional neural networks (CNN) and stacking multi-layer bi-directional long-term short memory (BiLSTM) for thermal prediction. The findings of the experiments reveal that the model successfully anticipates the temperature with the highest R2value of 97.15% and the lowest error rate of RMSE value of 0.2892, and an RMAE of 0.5003, which decreases the projection error as opposed to the other method.

This research uses molecular dynamics (MD) simulations to examine how spherical roughness barrier... more This research uses molecular dynamics (MD) simulations to examine how spherical roughness barriers affect the boiling behavior of argon particles in a microchannel with a square crosssection. To establish boiling conditions, constant wall temperatures ranging from 84K to 133K are applied. Results indicate that roughness elements aid the thermal force of boiling by enhancing the distribution of fluid atoms in the microchannel’s central layers, particularly at later time steps (750,000 to 1,000,000) and at higher wall temperatures of 114K and 133K. Summing velocity values revealed that while rough channel surfaces slightly reduce fluid velocity as wall temperature increases, velocity remains nearly unchanged in smooth channels. Statistical analysis shows that spherical roughness does not negatively impact argon boiling properties and, in fact, enhances the effective heat transfer surfaces, thereby strengthening boiling conditions. Consequently, polishing microchannel surfaces or removing spherical roughness may even be detrimental for certain practical applications.

Solar energy is one of the renewable and clean energy sources. Accurate solar radiation (SR) esti... more Solar energy is one of the renewable and clean energy sources. Accurate solar radiation (SR) estimates aretherefore needed in solar energy applications. Firstly, two deep learning models, including gated recurrent unit(GRU) and long short-term memory (LSTM), were developed in this study. Next, a data pre-processing techniquenamed multivariate variational mode decomposition (MVMD) was used to construct the MVMD-GRU andMVMD-LSTM hybrid models. To better test the performance of proposed simple and hybrid models, four stationslocated in the Illinois State of the USA (i.e., Dixon Springs, Fairfield, Rend Lake, and Carbondale) wereconsidered as the study sites. Whole the simple and hybrid models were established under two different strategies, i.e., local and external. In the local strategy, SR of each location was estimated using the minimum andmaximum air temperatures from the same station. While, minimum and maximum air temperatures as well as SRdata from the nearby station were utilized in external strategy to estimate SR time series of any target site. Rootmean square error (RMSE), mean absolute error (MAE), and coefficient of determination (R2) metrics were usedwhen evaluating the models performances. The overall results revealed that the proposed MVMD-GRU andMVMD-LSTM hybrid models illustrated better SR estimates compared to the simple GRU and LSTM in both thelocal and external strategies. The values of error metrics obtained for the superior hybrid models (i.e., MVMDLSTM) during the testing period were as: RMSE = 2.532 MJ/m2.day, MAE = 1.921 MJ/m2.day, R2 = 0.916 atDixon Springs; RMSE = 2.476 MJ/m2.day, MAE = 1.878 MJ/m2.day, R2 = 0.921 at Fairfield; RMSE = 2.359 MJ/m2.day, MAE = 1.780 MJ/m2.day, R2 = 0.924 at Rend Lake; RMSE = 2.576 MJ/m2.day, MAE = 1.941 MJ/m2.day, R2 = 0.914 at Carbondale. Therefore, the coupled models proposed in this study can be possibly recommended as suitable alternatives to the simple deep learning models with a reliable precision in estimating SRtime series

PRO PUBLICO BONO – Public Administration, 2024

This article presents a state-of-the-art review of machine learning (ML) methods and applications... more This article presents a state-of-the-art review of machine learning (ML) methods and applications used in smart grids to predict and optimise energy management. The article discusses the challenges facing smart grids, and how ML can help address them, using a new taxonomy to categorise ML models by method and domain. It describes the different ML techniques used in smart grids as well as examining various smart grid use cases, including demand response, energy forecasting, fault detection, and grid optimisation, and explores how ML can improve these cases. The article proposes a new taxonomy for categorising ML models and evaluates their performance based on accuracy, interpretability, and computational efficiency. Finally, it discusses some of the limitations and challenges of using ML in smart grid applications and attempts to predict future trends. Overall, the article highlights how ML can enable efficient and reliable smart grid systems.

PRO PUBLICO BONO – Public Administration,, 2024

Artificial intelligence (AI) is widely used in social sciences and continues to evolve. Deep lear... more Artificial intelligence (AI) is widely used in social sciences and continues to evolve. Deep learning(DL) has emerged as a powerful AI tool transforming social sciences with valuable insights across many areas. Employing DL for modelling social sciences’ big data has led to significant discoveries and transformations. This study aims to systematically review and evaluate DL methods in social sciences. Following PRISMA guideline, this study identifies fundamental DL methods applied to social science applications. We evaluated DL models using reported metrics and calculated normalized reliability score for uniform assessment. Employing relief feature selection,we identified influential parameters affecting DL techniques’ reliability. Findings suggest evaluation criteria significantly impact DL model effectiveness, while database and application type influence moderately. Identified limitations include inadequate reporting of evaluation criteria and model structure details hindering comprehensive assessment and informed policy development. In conclusion, this review underscores DL methods’ transformative role in social sciences, emphasizing the importance of explainability and responsibility.

Ground deformation, due to tunneling, is one of the most significant challenges in tunnel design ... more Ground deformation, due to tunneling, is one of the most significant challenges in tunnel design in soft ground along with, the predicting the related effects of tunneling on nearby structures. One of the methods of predicting ground settlement in tunneling projects, is to use analytical and numerical methods. By measuring the amount of settlement with accurate instruments and back-analysis of behavioral measurement data, in addition to estimating the state of settlement of the ground and surrounding structures, it is possible to determine the geotechnical parameters of the soil and structure in the design of upcoming sections and future designs. In this study, an attempt has been made to verify the measured settlements caused by digging the tunnel of an urban train line, by using back analysis. For this purpose, comparisons with predictions obtained from empirical and analytical methods and the Geotechnical Engineering Finite Element Analysis software (PLAXIS) was used. The results show that often, the empirical methods obtain values more than the measured values, for ground settlement.

Journal of King Saud University - Computer and Information Sciences , 2024

Machine learning contributes in improving the accuracy of melanoma detection. There are extensive... more Machine learning contributes in improving the accuracy of melanoma detection. There are extensive studies in classic and deep learning-based approaches for melanoma detection in the literature. Still, they are not accurate or require high learning data. This paper proposes a hybrid mechanism for automated melanoma detection on dermoscopic images based on Discrete Cosine Transform features and metadata. It is composed of three steps. First, extra information/artifacts are deleted; the remaining pixels are standardized for accurate processing. Second, the reliability of the mechanism is improved by the Radon transform, extra data is removed using the Top-hat filter, and the detection rate is increased by Discrete Wavelet Transform and Discrete Cosine Transform. Then, the number of features is reduced by Locality Sensitive Discriminant Analysis. The third step divides the images into learning and test ones to create image-based models using learning data. Finally, the best model is selected based on test data and improved by a metadata-based model. Simulation results show that the decision tree provides the most practical image-based model by improving accuracy and sensitivity. Besides, the comparison results demonstrate that our model improves the F-Value to detect melanoma superior to other approaches.

2023 IEEE 17th International Symposium on Applied Computational Intelligence and Informatics (SACI)

Journal of Clinical Care and Skills

Alexandria Engineering Journal

When exposed to fluid flow, elastically supported multiple square cylinders may experience either... more When exposed to fluid flow, elastically supported multiple square cylinders may experience either one or a combination of vortex-induced vibration, galloping, and wake-induced vibration phenomena. Due to these high-amplitude instabilities, it is necessary to utilize vibration control devices. The present paper studies the suppression of flow-induced vibration (FIV) of tandemarranged square-section cylinders, which can oscillate independently in the streamwise and transverse directions at low Reynolds numbers. The vibration reduction is achieved by directly applying opposing forces using an active FIV control system based on the intelligent proportional-integralderivative controller. In the present study, the fluid flow equations are calculated through the finite volume technique, by which the aerodynamic forces are attained. Next, based on the computed excitation forces, the motion equations of cylinders are solved within a user-defined function code. The numerical results show that the controller successfully suppresses the vibrations of the front and rear cylinders by a maximum of 94% and 92% at Re = 80 and by 98% and 97% at Re = 100. At Re = 230 and 250, the controller successfully diminishes the oscillation of the front cylinder by a maximum of 79% and 80%, respectively. The significant intensity of the front cylinder's wake makes the active control system unable to capably affect the vibrations of the downstream cylinder at Re = 230 and 250.

PRO PUBLICO BONO – Public Administration, 2024

PRO PUBLICO BONO – Public Administration,, 2024

Journal of King Saud University - Computer and Information Sciences , 2024

2023 IEEE 17th International Symposium on Applied Computational Intelligence and Informatics (SACI)

Journal of Clinical Care and Skills

Alexandria Engineering Journal

Learning ability in Robotics is acknowledged as one of the major challenges facing artificial int... more Learning ability in Robotics is acknowledged as one of the major challenges facing artificial intelligence. Although in the numerous areas within Robotics machine learning (ML) has long identified as a core technology, recently Robot learning, in particular, has been witnessing major challenges due to the theoretical advancement at the boundary between optimization and ML. In fact the integration of ML and optimization reported to be able to dramatically increase the decision-making quality and learning ability in decision systems. Here the novel integration of ML and optimization which can be applied to the complex and dynamic contexts of Robot learning is described. Furthermore with the aid of an educational Robotics kit the proposed methodology is evaluated. 1 Introduction Today learning has become a major part of the research in Robotics [1]. Machine learning (ML) algorithms in robotics in particular, within autonomous control and sensing, are being used to tackle difficult problems where large quantities of datasets are available which enable Robots to effectively teach themselves [2]. ML as a sub-field of computer science has evolved from the study of pattern recognition and computational learning theory [3]. Furthermore ML is considered as a field of study in artificial intelligence that gives computers the ability to learn from data [4]. To do so ML explores the development of models that can predict and learn from an available dataset [5]. Such models operate with the aid of algorithms capable of making data-driven predictions rather than following explicit codes [6]. Consequently ML is often used in a range of problems where designing precise algorithms is not practical. In this sense ML can replace the human expertise in information treatment [7]. For that matter ML provides the algorithmic tools for dealing with datasets and providing predictions. In fact ML tends to imitate human skills, which in most cases, act exceptional in identifying satisfactory solutions by theoretical or experience-based considerations [8].

Information systems (IS) have been recently recognized to be the major tools to be highly utilize... more Information systems (IS) have been recently recognized to be the major tools to be highly utilized in supporting the decision-makers in the enterprises worldwide. However despite of all the recent advancements in developing the rational tools of information and communication technologies (ICT) for decision-making, e.g. decision support systems (DSS) and business inteligence (BI), still intuition plays effective role in decision-making under uncertainty and big data. In fact in today’s globally competitive, uncertain and dynamic business environments, understanding the concept of intuition and systematically using it more than ever is considered to be vital in fuelling the creativity, making fast decisions, reacting appropriately to the dynamic market, and also governing the information technology (IT).
As the topic of intuition may be investigated from different perspectives there has been a demand for a multidisciplinary research on the topic. While the mechanism, success/failure ratio, marvels and flaws of intuition are still under debate, here our revision on the latest researches on psychology and neuroscience of creativity proves that intuition cannot be always trusted in leading to the optimal decisions. Yet the permanent solution to creative decision-making would be an integration of intuition and rational tools. Further this report, a methodology is accordingly proposed in dealing with decision-making tasks under uncertainty and big data. A case study in engineering design is then given to evaluate the effectiveness of the methodology.
Furthermore along with carrying out the case studies, which have been previously separately published, the concepts of business modelling, requirement specification, algorithms implementation and software testing are well practiced.

The aim of this book is to investigate the role of business models and to explore the potential o... more The aim of this book is to investigate the role of business models and to explore the potential of business modeling for sustainable development in particular. Further the focus would be narrower to the research on sustainable product development. Thus this book provides a broad revision to the literature on business models for sustainable development. Our literature review is mainly concerned with the business models which have been adapted by firms involved in product development or the researches related to product development. This book concludes that using business modeling in product development can empower sustainability and highly contribute in sustainable development.

Nowadays big-data analytics has become an important tool for different engineering fields. Its fl... more Nowadays big-data analytics has become an important tool for different engineering fields. Its flexibility allows a constant increasing scope for several applications. Among others, the advantages of this approach are: time optimization, real time decision making, modeling and prediction. With them, it has been possible to find more accurate and feasible solutions for current engineering problems. Moreover, there are several new big-data applications within the industry. In one hand, in fields like Telecommunications, Information Technology (IT), Industry 4.0, Mechanical Engineering and others, a huge amount of data is always generated and consequently, the question of how to manage it is still open. On the other hand, in fields such as Government, Business, Banking, Health, and Education, the pursue of novel applications is producing a significant amount of data as well. Hence, in this research is addressed the impact of big-data analytics in the previous mentioned fields. Furthermore, the work presents how these fields have adopted big-data analytics within their processes, and the advantages attached to the use of those techniques. Finally, this survey shows the future trends of big-data analytics and its technical challenges as well.

Multiobjective optimization method of modeFrontier software is becoming popular now a day. The re... more Multiobjective optimization method of modeFrontier software is becoming popular now a day. The reason is that it utilizes the available resources in an efficient manner and provide results in less duration of time. This thesis survey the algorithm and method of modeFrontier to solve optimization problem in different sectors like Energy,Manufacturing, Material, Transportation, Bioscience, Aerospace, Metal forming,Electrical engineering, Health & Foundry. This research is focused on multiobjective optimization methods of modeFrontier. It involves various optimization methods like Design optimization, Numerical analysis, Computational Fluid Dynamics, Evolution alalgorithm etc. These methods are compiled together in modeFrontier that provide an easy workflow for different modules and act as a building block for solving multipledecision issues of product modeling. This work demonstrates the increasing demand of this software in the upcoming years which is justified by graphical representation. This thesis is an art of survey for reviewing the recent and past achievement of multiobjective optimization, involved tools and methods. The data collected from previous studies demonstrate that this software has huge advantage for companies and academics to carry out optimization in product modeling.

Dr. Amir Mosavi, adattudós, jelenleg az Óbudai Egyetem Kandó Kálmán Villamosmérnöki Karán folytat... more Dr. Amir Mosavi, adattudós, jelenleg az Óbudai Egyetem Kandó Kálmán Villamosmérnöki Karán folytat kutatásokat az alacsony szennyezőanyag kibocsátású hibrid motorok fejlesztése területén Dr. Dineva Adrienn és Prof. Dr. Várkonyiné-Kóczy Annamária kutatókkal tudományos kollaboráció keretén belül. Dr. Mosavi a Budapesti Gazdasági Főiskola meghívott tanára. Informatikai-illetve környezetmérnöki szakokon szerzett diplomát, majd az Egyesült Királyságban és Kanadában folytatott további tanulmányokat. Számos utazást tett, több, mint tizenöt kollaborációs projektben vett rész az adattudományok és döntéselmélet területén kiemelkedő, széleskörben elismert egyetemeken. Több, mint 55 országban járt. Utazásai során a " Predict It " című könyvéhez gyűjtött ismereteket és vizsgálta a környezetvédelemmel kapcsolatos döntési mechanizmusokat különböző kulturális és geopolitikai hatások alatt. Dr. Mosavi elismert vezető fenntarthatósági tudós, Dr. Johanna Wanka német oktatási és kutatási miniszter kitüntetését vehette át a fenntartható fejlődésben elért kiemelkedő eredményeiért. A kiemelkedő fenntartható-fejlődés kutatók között tartják számon, elnyerte a Green Talent Díjat az általa kifejlesztett pontos környezetvédelmi előrejelző módszerért, amely sikerrel alklamazható a környezetvédelem-menedzsmentben és katasztrófa-csökkentési rendelkezések tervezésében. Vallja, hogy az emberiségnek szüksége van pontos előrejező-döntéstámogató eszközökre annak érdekében, hogy képesek legyenek vizsgálni és csökkenteni a klímaváltozás hatásait. Dr. Mosavi nagymértékebn hozzájárult az éghajlatváltozás kockázatának csökkentését célzó kutatásokhoz az ún. prediktív döntési modell beveztésével. Az adat-vezérelt számítási platformja világszerte a legfejlettebb tanácsadást biztosítja a döntéshozók számára, hogy a legmagasabb tájékozottsági szinten hozzanak döntést. A lehetséges döntések következményeinek pontos előrejelzése a prediktív-döntési modell alkalmazása révén optimális/jobb és automatizált módon történik. A klímaváltozás valóságában a pontenciális döntéseket támogató fejlett eszközök valóban létfontosságúak, mivel a környezetbarát döntések gyakran ellentmondanak a nemzetek gazdasági és politikai érdekeinek. Az alapötlete tehát, hogy támogassa a szervezeteket és kormányokat olyan új üzleti modellek bevezetésével, amelyek fenntarthatók és sok generáció számára fenntartják jövedelmezőségüket. Ez az új koncepció különösen hasznos segítséget nyújt az UNESCO Bioszféra rezervátumainak fenntarhatósági kérdéseiben. Az adattudományok segítségével a múltbeli eseményeket elemezzük annak érdekében, hogy a jövőbeli döntésekről és cselekvésekről képet kapjunk. Tény, hogy a prediktív modellek segítségével a tudósok tanulmányozhatják az emberi viselkedés lehetséges hatásait, hogy képesek legyenek fenntarthatóbb irányba fejlődni és növekedni. Ez az újdonság forradalmasította napjaink döntéshozatali rendszereinek működését. Számos díjat, köztük a MAB UNESCO fiatal tudósok díját, a TU-Darmstadt Jövő tehetség-díját, a Graz-i GoStyria-díjat, a weimari Bauhaus Egyetem kutatói ösztöndíját, a Campus France Research díjat és az Észt Dora & Estophilus díjat vehette át Dr. Mosavi a meglepően újszerű és tudományosan szigorú előrejelzési modellekkel végzett kutatási eredményeiért, amelyek új utakat nyitnak és figyelemre méltó eredményekkel emelkednek ki. Az éghajlatváltozáshoz való alkalmazkodáshoz és a kockázatcsökkentéshez javasolt adatközpontú döntéstámogató rendszere valóban kiterjedt lehetőség arra, hogy a legfrissebb felfedezéseket a döntési tudomány és előrejelzés egyik fő innovációjává alakítsa. A közelmúltban a Norvég Tudomány és Technológia Egyetem az Informatika és Matematika Európai Kutatási Konzorciuma kutatási alapján keresztül támogatta Dr. Mosavi kutatómunkáját, ahol az innovatív koncepciót a következő szintre emelheti gyümölcsöző együttműködéssel vezető tudósokkal, mint Pinar Ozturk és Hai Thanh Nguyen professzorok.

Dr Amir Mosavi receives the UNESCO Young Scientists Award The International Coordinating Council ... more Dr Amir Mosavi receives the UNESCO Young Scientists Award
The International Coordinating Council of the Man and the Biosphere (MAB) Programme of UNESCO has awarded Dr. Amir Mosavi with the Young Scientists Award in Paris, France on 15 June.

The recent developments of computer and electronic systems have made the use of intelligent syste... more The recent developments of computer and electronic systems have made the use of intelligent systems for the automation of agricultural industries. In this study, the temperature variation of the mushroom growing room was modeled by multi-layered perceptron and radial basis function networks based on independent parameters including ambient temperature, water temperature, fresh air and circulation air dampers, and water tap. According to the obtained results from the networks, the best network for MLP was in the second repetition with 12 neurons in the hidden layer and in 20 neurons in the hidden layer for radial basis function network. The obtained results from comparative parameters for two networks showed the highest correlation coefficient (0.966), the lowest root mean square error (RMSE) (0.787) and the lowest mean absolute error (MAE) (0.02746) for radial basis function. Therefore, the neural network with radial basis function was selected as a predictor of the behavior of the system for the temperature of mushroom growing halls controlling system.

Prediction of crops yield is essential for food security policymaking, planning, and trade. The o... more Prediction of crops yield is essential for food security policymaking, planning, and trade. The objective of the current study is to propose novel crop yield prediction models based on hybrid machine learning methods. In this study the performance of artificial neural networks-imperialist competitive algorithm (ANN-ICA) and artificial neural networks-gray wolf optimizer (ANN-GWO) models for the crop yield prediction are evaluated. According to the results, ANN-GWO, with R of 0.48, RMSE of 3.19, and MEA of 26.65, proved a better performance in the crop yield prediction compared to the ANN-ICA model. The results can be used by either practitioners, researchers or policymakers for food security.

One of the most common and important destructive attacks on the victim system is Advanced Persist... more One of the most common and important destructive attacks on the victim system is Advanced Persistent Threat (APT)-attack. The APT attacker can achieve his hostile goals by obtaining information and gaining financial benefits regarding the infrastructure of a network. One of the solutions to detect a secret APT attack is using network traffic. Due to the nature of the APT attack in terms of being on the network for a long time and the fact that the network may crash because of high traffic, it is difficult to detect this type of attack. Hence, in this study, machine learning methods such as C5.0 decision tree, Bayesian network and deep neural network are used for timely detection and classification of APT-attacks on the NSL-KDD dataset. Moreover, 10-fold cross validation method is used to experiment these models. As a result, the accuracy (ACC) of the C5.0 decision tree, Bayesian network and 6-layer deep learning models is obtained as 95.64%, 88.37% and 98.85%, respectively, and also, in terms of the important criterion of the false positive rate (FPR), the FPR value for the C5.0 decision tree, Bayesian network and 6-layer deep learning models is obtained as 2.56, 10.47 and 1.13, respectively. Other criterions such as sensitivity, specificity, accuracy, false negative rate and F-measure are also investigated for the models, and the experimental results show that the deep learning model with automatic multi-layered extraction of features has the best performance for timely detection of an APT-attack comparing to other classification models.

The current study assessed the influence of Anti Stripping Agents (ASA), Ground Tire Rubber (GTR)... more The current study assessed the influence of Anti Stripping Agents (ASA), Ground Tire Rubber (GTR) and waste polyethylene terephthalate (PET) on performance behavior of binder and Stone Matrix Asphalt (SMA) mixtures. Through this paper the 85/100 penetration grade bitumen was utilized as original bitumen. Also, three liquid ASA’s (ASA (A), ASA (B), ASA (C)) were used as mixture modifier. For this purpose, softening point, penetration, rotational viscosity, Dynamic Shear Rheometer, Multi Stress Creep Recovery (MSCR) and Linear Amplitude Sweep (LAS) tests were implemented to investigate the rheological properties of modified bitumen. For evaluating the behavior of modified mixtures several tests such as; Resilient Modulus, Tensile Strength, dynamic creep, wheel track and four point beam fatigue tests were implemented. Based on MSCR test results, utilization of mentioned polymers enhanced the elasticity of bitumens and therefore the permanent deformation resistance of binders increases. Also by addition of PET percentage, the rutting resistance improves. Results indicated that utilization of ASAs, PET and Crumb Rubber (CR) enhance the Resilient Modulus (Mr), Indirect Tensile Strength (ITS), rutting resistance, fatigue life and Fracture Energy (FE) of asphalt mixtures. Also based on results, modification of binder by PET/CR with ratio of 50%/50% and ASA (B) have the highest fatigue life which indicates that this mixture have highest resistance against fatigue cracking.