Emiliano Mankolli | Polis University, Tirana (original) (raw)
Uploads
Papers by Emiliano Mankolli
This paper aims to examine the existing vulnerabilities in the cybersecurity of a university camp... more This paper aims to examine the existing vulnerabilities in the cybersecurity of a university campus and the approaches and ways to prevent losses caused by wrong decisions due to the analysis of corrupted data. The data analysis is usually performed using machine learning (ML) algorithms and artificial intelligence (AI). An approach to avoiding AI and ML vulnerabilities is presented in this paper. Ways to prevent wrong conclusions in data analysis and to avoid making wrong decisions are discussed. Another important aspect of improving security is ensuring a high level of cyber security and preventing (hacker) attacks, which can greatly endanger information processes and communication in the university and can also lead to the occurrence of material damage and human victims. Examples are provided to show how massive cyberattacks are in higher schools, based on data from universities in the United States of America. Several methods of attack prevention and detection are presented from the perspective of machine learning applications.
2023 24th International Conference on Control Systems and Computer Science (CSCS)
This paper presents machine learning techniques used to predict a candidate's success for... more This paper presents machine learning techniques used to predict a candidate's success for a job vacancy. This new method for selecting candidates, supported by the use of multiple parameters, guarantees high efficiency and accuracy. Our approach shows that applying natural language processing methods to textual data related to the candidate's profile and the vacancy description significantly increases the overall accuracy. Boosted trees as implemented in XGBoost achieved the highest accuracy.
2022 29th International Conference on Systems, Signals and Image Processing (IWSSIP)
The recruitment industry is facing major challenges in terms of selecting candidates who best fit... more The recruitment industry is facing major challenges in terms of selecting candidates who best fit a vacancy. Artificial Intelligence, Machine Learning, and Natural Language Processing, although they have greatly increased the processing of mass data for candidates, are not without limitations. This paper presents a hybrid solution to candidate selection with a job title that is fitting for an open vacancy. This hybrid approach is a combination of two Machine Learning techniques, XGBoost and BERTopic. This solution reduces processing time and memory usage. The definition of job title similarity in the context of similar industries provides higher accuracy to the complex and delicate process such as candidate selection fitting for an open vacancy.
2023 IEEE INTERNATIONAL WORKSHOP ON Technologies for Defense and Security, 2023
This paper aims to examine the existing vulnerabilities in the cybersecurity of a university camp... more This paper aims to examine the existing vulnerabilities in the cybersecurity of a university campus and the approaches and ways to prevent losses caused by wrong decisions due to the analysis of corrupted data. The data analysis is usually performed using machine learning (ML) algorithms and artificial intelligence (AI). An approach to avoiding AI and ML vulnerabilities is presented in this paper. Ways to prevent wrong conclusions in data analysis and to avoid making wrong decisions are discussed. Another important aspect of improving security is ensuring a high level of cyber security and preventing (hacker) attacks, which can greatly endanger information processes and communication in the university and can also lead to the occurrence of material damage and human victims. Examples are provided to show how massive cyberattacks are in higher schools, based on data from universities in the United States of America. Several methods of attack prevention and detection are presented from the perspective of machine learning applications.
Proc. 29th International Conference on Systems, Signals and Image Processing “IWSSIP 2022”, 2022
The recruitment industry is facing major challenges in terms of selecting candidates who best fit... more The recruitment industry is facing major challenges in terms of selecting candidates who best fit a vacancy. Artificial Intelligence, Machine Learning, and Natural Language Processing, although they have greatly increased the processing of mass data for candidates, are not without limitations. This paper presents a hybrid solution to candidate selection with a job title that is fitting for an open vacancy. This hybrid approach is a combination of two Machine Learning techniques, XGBoost and BERTopic. This solution reduces processing time and memory usage. The definition of job title similarity in the context of similar industries provides higher accuracy to the complex and delicate process such as candidate selection fitting for an open vacancy.
International Conference on Control Systems and Computer Science (CSCS), 2023
This paper presents machine learning techniques used to predict a candidate's success for a job v... more This paper presents machine learning techniques used to predict a candidate's success for a job vacancy. This new method for selecting candidates, supported by the use of multiple parameters, guarantees high efficiency and accuracy. Our approach shows that applying natural language processing methods to textual data related to the candidate's profile and the vacancy description significantly increases the overall accuracy. Boosted trees as implemented in XGBoost achieved the highest accuracy.
2021 15th International Conference on Advanced Technologies, Systems and Services in Telecommunications (TELSIKS), 2021
This paper introduces a hybrid method based on the combination of two machine learning methods (k... more This paper introduces a hybrid method based on the combination of two machine learning methods (k-NN and SVM). The novel method is designed to find job titles that are similar based on their description and industry. This method improves the accuracy and time-efficiency of a complex process like selecting the best candidates for a job.
Communications in Computer and Information Science
The purpose of this paper is to consider one of the most important modern technologies, namely: N... more The purpose of this paper is to consider one of the most important modern technologies, namely: Natural Language Processing (NLP) and the machine learning algorithms related to it. The aim of the authors is to present the machine learning models in this interdisciplinary scientific field, which represents an intersection point of computer science, artificial intelligence, and linguistics. The machine learning techniques are classified and the corresponding models are briefly discussed. Different optimization approaches and problems for machine learning are considered. In this regard, some conclusions are drawn about the development trends in the area and the directions for future research.
This paper aims to examine the existing vulnerabilities in the cybersecurity of a university camp... more This paper aims to examine the existing vulnerabilities in the cybersecurity of a university campus and the approaches and ways to prevent losses caused by wrong decisions due to the analysis of corrupted data. The data analysis is usually performed using machine learning (ML) algorithms and artificial intelligence (AI). An approach to avoiding AI and ML vulnerabilities is presented in this paper. Ways to prevent wrong conclusions in data analysis and to avoid making wrong decisions are discussed. Another important aspect of improving security is ensuring a high level of cyber security and preventing (hacker) attacks, which can greatly endanger information processes and communication in the university and can also lead to the occurrence of material damage and human victims. Examples are provided to show how massive cyberattacks are in higher schools, based on data from universities in the United States of America. Several methods of attack prevention and detection are presented from the perspective of machine learning applications.
2023 24th International Conference on Control Systems and Computer Science (CSCS)
This paper presents machine learning techniques used to predict a candidate's success for... more This paper presents machine learning techniques used to predict a candidate's success for a job vacancy. This new method for selecting candidates, supported by the use of multiple parameters, guarantees high efficiency and accuracy. Our approach shows that applying natural language processing methods to textual data related to the candidate's profile and the vacancy description significantly increases the overall accuracy. Boosted trees as implemented in XGBoost achieved the highest accuracy.
2022 29th International Conference on Systems, Signals and Image Processing (IWSSIP)
The recruitment industry is facing major challenges in terms of selecting candidates who best fit... more The recruitment industry is facing major challenges in terms of selecting candidates who best fit a vacancy. Artificial Intelligence, Machine Learning, and Natural Language Processing, although they have greatly increased the processing of mass data for candidates, are not without limitations. This paper presents a hybrid solution to candidate selection with a job title that is fitting for an open vacancy. This hybrid approach is a combination of two Machine Learning techniques, XGBoost and BERTopic. This solution reduces processing time and memory usage. The definition of job title similarity in the context of similar industries provides higher accuracy to the complex and delicate process such as candidate selection fitting for an open vacancy.
2023 IEEE INTERNATIONAL WORKSHOP ON Technologies for Defense and Security, 2023
This paper aims to examine the existing vulnerabilities in the cybersecurity of a university camp... more This paper aims to examine the existing vulnerabilities in the cybersecurity of a university campus and the approaches and ways to prevent losses caused by wrong decisions due to the analysis of corrupted data. The data analysis is usually performed using machine learning (ML) algorithms and artificial intelligence (AI). An approach to avoiding AI and ML vulnerabilities is presented in this paper. Ways to prevent wrong conclusions in data analysis and to avoid making wrong decisions are discussed. Another important aspect of improving security is ensuring a high level of cyber security and preventing (hacker) attacks, which can greatly endanger information processes and communication in the university and can also lead to the occurrence of material damage and human victims. Examples are provided to show how massive cyberattacks are in higher schools, based on data from universities in the United States of America. Several methods of attack prevention and detection are presented from the perspective of machine learning applications.
Proc. 29th International Conference on Systems, Signals and Image Processing “IWSSIP 2022”, 2022
The recruitment industry is facing major challenges in terms of selecting candidates who best fit... more The recruitment industry is facing major challenges in terms of selecting candidates who best fit a vacancy. Artificial Intelligence, Machine Learning, and Natural Language Processing, although they have greatly increased the processing of mass data for candidates, are not without limitations. This paper presents a hybrid solution to candidate selection with a job title that is fitting for an open vacancy. This hybrid approach is a combination of two Machine Learning techniques, XGBoost and BERTopic. This solution reduces processing time and memory usage. The definition of job title similarity in the context of similar industries provides higher accuracy to the complex and delicate process such as candidate selection fitting for an open vacancy.
International Conference on Control Systems and Computer Science (CSCS), 2023
This paper presents machine learning techniques used to predict a candidate's success for a job v... more This paper presents machine learning techniques used to predict a candidate's success for a job vacancy. This new method for selecting candidates, supported by the use of multiple parameters, guarantees high efficiency and accuracy. Our approach shows that applying natural language processing methods to textual data related to the candidate's profile and the vacancy description significantly increases the overall accuracy. Boosted trees as implemented in XGBoost achieved the highest accuracy.
2021 15th International Conference on Advanced Technologies, Systems and Services in Telecommunications (TELSIKS), 2021
This paper introduces a hybrid method based on the combination of two machine learning methods (k... more This paper introduces a hybrid method based on the combination of two machine learning methods (k-NN and SVM). The novel method is designed to find job titles that are similar based on their description and industry. This method improves the accuracy and time-efficiency of a complex process like selecting the best candidates for a job.
Communications in Computer and Information Science
The purpose of this paper is to consider one of the most important modern technologies, namely: N... more The purpose of this paper is to consider one of the most important modern technologies, namely: Natural Language Processing (NLP) and the machine learning algorithms related to it. The aim of the authors is to present the machine learning models in this interdisciplinary scientific field, which represents an intersection point of computer science, artificial intelligence, and linguistics. The machine learning techniques are classified and the corresponding models are briefly discussed. Different optimization approaches and problems for machine learning are considered. In this regard, some conclusions are drawn about the development trends in the area and the directions for future research.