Esref ADALI | Istanbul Technical University (original) (raw)

Papers by Esref ADALI

Türkiye Bilişim Derneği (TBD), 2020

Türkiye Bilişim Vakfı Bilgisayar Bilimleri ve Mühendisliği Dergisi, Jun 24, 2016

Türkiye Bilişim Vakfı Bilgisayar Bilimleri ve Mühendisliği Dergisi, Dec 17, 2019

Question and answering system use either own knowledge base or data collected from Internet. In t... more Question and answering system use either own knowledge base or data collected from Internet. In this paper a knowledge-based question and answering system for Turkish is presented. Turkish is member of Altay language family and an agglutinative language. The grammar of Turkish different then India-European languages. Therefore, this project had been developed according to Turkish grammar. System consists of several steps. These steps are: preprocessing (tokenizing of sentences, parsing a word into syllable, phonological testing (consonant softening, vowel reduction, haplology, head and tail letters), morphological testing, mumpsimus, foreign words); morphological parsing; finding of roles of words; interpretation of question and answering. In the first phase the QA system can understand question and answers. Later on, constitutes question and answer parallel corpus in other word data bases. The second phase consist of the same pre-processing then searches for the same or similar question in question database if finds the answer gets from answer database. In order to fine similar question similarity vectors approaches were used.

Published for the International Federation of Automatic Control by Pergamon Press eBooks, 1987

İTÜ Dergisi D : Mühendislik, 2008

... Türkmenceden Türkçeye bilgisayarlı metin çevirisi Ahmet Cüneyd TANTUĞ*1, Eşref ADALI1, Kemal ... more ... Türkmenceden Türkçeye bilgisayarlı metin çevirisi Ahmet Cüneyd TANTUĞ*1, Eşref ADALI1, Kemal OFLAZER2 ... İstatiksel dil modeli Çeviri sisteminin bileşenleri içerisinde, Türk-mence biçimbilimsel çözümleyici ve kök akta-rımı bileşenlerinin çıktıları belirsizlik içermektedir: 89 ...

CRC Press eBooks, Feb 4, 2022

Türkiye Bilişim Vakfı Bilgisayar Bilimleri ve Mühendisliği Dergisi, 2016

In the study, it is aimed to run a large-scale application, which is suitable for parallelization... more In the study, it is aimed to run a large-scale application, which is suitable for parallelization\ at the embedded computers with using central management software that taking computer cluster architecture as reference. Pool backup approach is used to increase system reliability. On the other hand, system efficiency was increased by (1) optimizing file transfers by avoiding redundant transfers via MD5 based controls, (2) using the scoring algorithm designed as part of this study to prioritize the nodes providing a mechanism to assign priority scores to subtasks to be executed at nodes, (3) homogenizing use among all nodes to equalize weathering rates and delay the first breakdown.

Degaussing is a method used to decreasing ships' magnetic field in order to minimize the poss... more Degaussing is a method used to decreasing ships' magnetic field in order to minimize the possibility of detection by magnetic influenced mechanisms. Degaussing ranges serve to control the magnetic fields around ship and to decrease it. This paper is dealing with a modernization project of a degaussing range which belongs to the Turkish Navy in Istanbul. The modernization project of the degaussing range house is based on today's high technologies. All old components of the range house such as fluxmeters, control panel has been replaced with a modem, computer based system (CBDR). CBDR provides new hardware and various software facilities for degaussing range. Because of the hardware and software modularity of the CBDR, it can be easily expanded for the future requirements. During the modernization of the former system, the existent coil ranges have been kept same. Because of using the same coil ranges, the theoretical principle and formulations of the former system have not been changed in the developed system. The same test techniques and the formulas have been used. The main objectives to design such a project are to eliminate human factor, to obtain more precise measurement, to decrease the cost of system in terms of investment and operation, to provide an intelligent filing system for collected test results and degaussing knowledge about the ships by using a relational data base management system, to provide a faster calibration process for regular ship by using an optimal current searching algorithm, to facilitate tests by providing user friendly and robust operator interface software. The software developed for this project consists of eight main modules which are Operator Interface, Data Collection and Evaluation, Graphics, Calculation and Evaluation, Optimal Degaussing Currents Search, Relational Data Base Management, Reporting and Statistics, and Maintenance modules.

Calismada, paralellestirmeye uygun olan bir uygulamanin bilgisayar kumeleri mimarisi referans ali... more Calismada, paralellestirmeye uygun olan bir uygulamanin bilgisayar kumeleri mimarisi referans alinarak gelistirilen merkez yonetim yazilimi vasitasiyla atanmis bilgisayarlarda calistirilmasi gerceklestirilmistir. Guvenilirligin artirilmasina yonelik olarak havuz yedekleme yontemi kullanilmistir. Diger taraftan verimi artirmaya yonelik olarak ise, (1) dosya aktarimlarinda, MD5 bazli kontroller konularak dosya aktarimi optimize edilmis ve ayni dosyalarin tekrarli aktarimi engellenmis, (2) tasarlanan onceliklendirme algoritmasi ile dugumler onceliklendirilmis, boylece dugumlerde calistirilacak uygulama alt gorevlerine oncelik degeri atama imkani verilmis, (3) dugumlerin homojen kullanimi saglanarak yaslanma oranlari dengelenmis, boylece sistemin ilk bozulma ani otelenmistir. Sistemin sinanmasi iki farkli bilgisayar ve dort farkli atanmis bilgisayar kullanarak gerceklestirilmistir. Sinanan atanmis bilgisayarlar sirasiyla, Raspberry PI Model B, Raspberry PI 2 Model B, BeagleBone Black M...

Gunumuzde akilli kartlar siklikla bilgi teknolojileri alaninda kullanilmaya basladigindan guvenli... more Gunumuzde akilli kartlar siklikla bilgi teknolojileri alaninda kullanilmaya basladigindan guvenlik onemli bir konu haline gelmistir. Bu yazida ozellikle akilli kartlarin guvenligine yonelik saldirilar anlatilmistir. Akilli kart gelistirilmesi ile ilgili calismalar sirasinda dikkat edilmesi gereken guvenlik konulari uzerinde durulmustur. Bu makale guvenli akilli kart isletim sistemi gereklerini ve akilli kart donanim ve isletim sisteminin saglamasi gereken ortak olcut guvence seviyesini belirtir. Ayrica bu makalede “dinamik guvenlik islevi” olarak adlandirilan yeni guvenlik yaklasimi ve kullanimi da anlatilmistir. : Abstract Today, as the smart cards are frequently used in Information Technologies, smart card security has become an important issue. In this paper, security attacks that are specific to smart cards are described. Security considerations that need to be taken into account when developing a secure smart card are studied. The paper indicates that the requirements of the se...

Soru yanitlama dizgeleri uzerinde surdurulen calismalar ya bilgi tabanli ya da genel agdaki veril... more Soru yanitlama dizgeleri uzerinde surdurulen calismalar ya bilgi tabanli ya da genel agdaki verileri kullanarak calismaktadir. Bu makalede tanitilan calisma bilgi tabanlidir ve Turkceye ozgudur. Ural dil ailesinin uyesi olan Turkce sondan eklemeli bir dildir ve dil bilgisi kurallari bukunlu Hint-Avrupa dillerinin dil bilgisi kurallarindan farklidir. Bu nedenle, Turkce yanitlama dizgesi Turkcenin dil bilgisi kurallarina uygun olarak hazirlanmistir. Dizge sirasiyla su asamalardan olusmaktadir: On isleme (tumcelere ayirma, sozcuklere ayirma, hecelere ayirma, ses bilimi acisindan sinama (unlu, unsuz uyumu, unlu dusmesi, bas ve sondaki harf kurallari, galatlar ve yabanci sozcukler); Bicim bilimsel cozumleme; Sozcuklerin siniflarinin belirlenmesi; Soz obeklerinin bulunmasi; Sozcuklerin rollerinin belirlenmesi; Tumcenin anlamlandirilmasi; Uygun yanitin uretilmesi. Uretilen her yanitin dogrulugu soru soran kisiye onaylatilmaktadir. Sorulan sorular ve bunlara iliskin yanitlar bir kosut derle...

Morphological units carry vast amount of semantic information for languages with rich inflectiona... more Morphological units carry vast amount of semantic information for languages with rich inflectional and derivational morphology. In this paper we show how morphosemantic information available for morphologically rich languages can be used to reduce manual effort in creating semantic resources like PropBank and VerbNet; to increase performance of word sense disambiguation, semantic role labeling and related tasks. We test the consistency of these features in a pilot study for Turkish and show that; 1) Case markers are related with semantic roles and 2) Morphemes that change the valency of the verb follow a predictable pattern.

Theory and applications of natural language processing, 2018

We present an approach to MT between Turkic languages and present results from an implementation ... more We present an approach to MT between Turkic languages and present results from an implementation of a MT system from Turkmen to Turkish. Our approach relies on ambiguous lexical and morphological transfer augmented with target side rule-based repairs and rescoring with statistical language models.

2018 3rd International Conference on Computer Science and Engineering (UBMK), Sep 1, 2018

Law enforcement authorities continually confront new and vexatious frauds involving computer viru... more Law enforcement authorities continually confront new and vexatious frauds involving computer virus attacks or credit card frauds committed against financial institutions and banking card companies. Such phenomenon used to commit online banking fraudulent transactions are occurring continuously and lead to a loss of gargantuan amount of money. Information security specialists and associations usually refer to the process of selecting practitioners as safeguards of fraud assessment, fraud analysis, or fraud detection. Designating such safeguards based on fraud calculations can be a exasperating and high-price process over the past few decades. This paper aims to understand how deep learning (DL) models can be benevolent in detecting fraudulent transactions with high accuracy. Dataset is extracted from one month of genuine financial logs of a mobile money service company in Africa, containing about more than six million transactions. The most satisfactory machine learning techniques as ensemble of decision tree (EDT), and deep learning techniques as stacked auto-encoders (SAE) and Restricted Boltzmann Machines (RBM) classifiers are applied on the preprocessed data. The performance of generated classifier models is evaluated based on accuracy, sensitivity, specificity, precision, confusion matrix and ROC values. The results of optimal accuracy are 90.49%, 80.52% and 91.53% respectively. The comparative results manifest that restricted Boltzmann machine performs superior than the other techniques.

2017 International Conference on Computer Science and Engineering (UBMK), Oct 1, 2017

Fraud detection is an enduring topic that pose a threat to banking, insurance, financial sectors ... more Fraud detection is an enduring topic that pose a threat to banking, insurance, financial sectors and information security systems such as intrusion detection systems (IDS), etc. Data mining and machine learning techniques help to anticipate and quickly detect fraud and take immediate action to minimize costs. This paper starts with the definition of intrusion detection system and its types, focuses on the implementation of a set of well-known machine learning classification algorithms (Decision Trees, Naive Bayes and Artificial Neural Networks), which can reduce the existing disadvantages of the intrusion detection systems. Experimental results on NSL-KDD dataset infer that our ANN-MLP method (Multilayer Perceptron) yields average better performance by calculating “confusion matrix” that in turn helps us to calculate performance measure such as, “Detection Rate Accuracy”, “precision” and “recall”.

Springer eBooks, 2006

This paper describes the implementation of a two-level morphological analyzer for the Turkmen Lan... more This paper describes the implementation of a two-level morphological analyzer for the Turkmen Language. Like all Turkic languages, the Turkmen Language is an agglutinative language that has productive inflectional and derivational suffixes. In this work, we implemented a finite-state two-level morphological analyzer for Turkmen Language by using Xerox Finite State Tools.

2017 International Conference on Computer Science and Engineering (UBMK), Oct 1, 2017

An intrusion detection system (IDS) monitors the network traffic looking for suspicious or malici... more An intrusion detection system (IDS) monitors the network traffic looking for suspicious or malicious activities or policy violations, which could represent an attack or unauthorized access. Traditional systems were designed to detect known attacks but cannot identify unknown threats. They most commonly detect known threats based on predefined rules or behavioral analysis through baselining the network. Expert attackers can bypass these techniques, therefore the need for more intelligent intrusion detection is increasing by the day. Intelligent IDS collects and analyzes information from different areas and deals with large amount of data, which contains various redundant and irrelevant features and results in increased overfitting, processing (training) time and low detection rate. For that reason, feature selection should be applied as a preprocessing step to improve accuracy and system performance while shrinking large data sets. In this paper, filter method which is a method used in feature selection to identify important features and to eliminate the less effective features has been developed for three different class structures on the NSL KDD dataset. After this process, classification models were created and performance measures were used to evaluate the models.

Language Resources and Evaluation, May 4, 2017

In this work, we report large-scale semantic role annotation of arguments in the Turkish dependen... more In this work, we report large-scale semantic role annotation of arguments in the Turkish dependency treebank, and present the first comprehensive Turkish semantic role labeling (SRL) resource: Turkish Proposition Bank (PropBank). We present our annotation workflow that harnesses crowd intelligence, and discuss the procedures for ensuring annotation consistency and quality control. Our discussion focuses on syntactic variations in realization of predicate-argument structures, and the large lexicon problem caused by complex derivational morphology. We describe our approach that exploits framesets of root verbs to abstract away from syntax and increase self-consistency of the Turkish PropBank. The issues that arise in the annotation of verbs derived via valency changing morphemes, verbal nominals, and nominal verbs are explored, and evaluation results for inter-annotator agreement are provided. Furthermore, semantic layer described here is aligned with universal dependency (UD) compliant treebank and released to enable more researchers to work on the problem. Finally, we use PropBank to establish a baseline score of 79.10 F1 for Turkish SRL using the mate-tool (an open-source SRL tool based on supervised machine learning) enhanced with basic morphological features. Turkish PropBank and the extended SRL system are made publicly available.

2022 7th International Conference on Computer Science and Engineering (UBMK), Sep 14, 2022

Türkiye Bilişim Derneği (TBD), 2020

Türkiye Bilişim Vakfı Bilgisayar Bilimleri ve Mühendisliği Dergisi, Jun 24, 2016

Türkiye Bilişim Vakfı Bilgisayar Bilimleri ve Mühendisliği Dergisi, Dec 17, 2019

Question and answering system use either own knowledge base or data collected from Internet. In t... more Question and answering system use either own knowledge base or data collected from Internet. In this paper a knowledge-based question and answering system for Turkish is presented. Turkish is member of Altay language family and an agglutinative language. The grammar of Turkish different then India-European languages. Therefore, this project had been developed according to Turkish grammar. System consists of several steps. These steps are: preprocessing (tokenizing of sentences, parsing a word into syllable, phonological testing (consonant softening, vowel reduction, haplology, head and tail letters), morphological testing, mumpsimus, foreign words); morphological parsing; finding of roles of words; interpretation of question and answering. In the first phase the QA system can understand question and answers. Later on, constitutes question and answer parallel corpus in other word data bases. The second phase consist of the same pre-processing then searches for the same or similar question in question database if finds the answer gets from answer database. In order to fine similar question similarity vectors approaches were used.

Published for the International Federation of Automatic Control by Pergamon Press eBooks, 1987

İTÜ Dergisi D : Mühendislik, 2008

... Türkmenceden Türkçeye bilgisayarlı metin çevirisi Ahmet Cüneyd TANTUĞ*1, Eşref ADALI1, Kemal ... more ... Türkmenceden Türkçeye bilgisayarlı metin çevirisi Ahmet Cüneyd TANTUĞ*1, Eşref ADALI1, Kemal OFLAZER2 ... İstatiksel dil modeli Çeviri sisteminin bileşenleri içerisinde, Türk-mence biçimbilimsel çözümleyici ve kök akta-rımı bileşenlerinin çıktıları belirsizlik içermektedir: 89 ...

CRC Press eBooks, Feb 4, 2022

Türkiye Bilişim Vakfı Bilgisayar Bilimleri ve Mühendisliği Dergisi, 2016

In the study, it is aimed to run a large-scale application, which is suitable for parallelization... more In the study, it is aimed to run a large-scale application, which is suitable for parallelization\ at the embedded computers with using central management software that taking computer cluster architecture as reference. Pool backup approach is used to increase system reliability. On the other hand, system efficiency was increased by (1) optimizing file transfers by avoiding redundant transfers via MD5 based controls, (2) using the scoring algorithm designed as part of this study to prioritize the nodes providing a mechanism to assign priority scores to subtasks to be executed at nodes, (3) homogenizing use among all nodes to equalize weathering rates and delay the first breakdown.

Degaussing is a method used to decreasing ships' magnetic field in order to minimize the poss... more Degaussing is a method used to decreasing ships' magnetic field in order to minimize the possibility of detection by magnetic influenced mechanisms. Degaussing ranges serve to control the magnetic fields around ship and to decrease it. This paper is dealing with a modernization project of a degaussing range which belongs to the Turkish Navy in Istanbul. The modernization project of the degaussing range house is based on today's high technologies. All old components of the range house such as fluxmeters, control panel has been replaced with a modem, computer based system (CBDR). CBDR provides new hardware and various software facilities for degaussing range. Because of the hardware and software modularity of the CBDR, it can be easily expanded for the future requirements. During the modernization of the former system, the existent coil ranges have been kept same. Because of using the same coil ranges, the theoretical principle and formulations of the former system have not been changed in the developed system. The same test techniques and the formulas have been used. The main objectives to design such a project are to eliminate human factor, to obtain more precise measurement, to decrease the cost of system in terms of investment and operation, to provide an intelligent filing system for collected test results and degaussing knowledge about the ships by using a relational data base management system, to provide a faster calibration process for regular ship by using an optimal current searching algorithm, to facilitate tests by providing user friendly and robust operator interface software. The software developed for this project consists of eight main modules which are Operator Interface, Data Collection and Evaluation, Graphics, Calculation and Evaluation, Optimal Degaussing Currents Search, Relational Data Base Management, Reporting and Statistics, and Maintenance modules.

Calismada, paralellestirmeye uygun olan bir uygulamanin bilgisayar kumeleri mimarisi referans ali... more Calismada, paralellestirmeye uygun olan bir uygulamanin bilgisayar kumeleri mimarisi referans alinarak gelistirilen merkez yonetim yazilimi vasitasiyla atanmis bilgisayarlarda calistirilmasi gerceklestirilmistir. Guvenilirligin artirilmasina yonelik olarak havuz yedekleme yontemi kullanilmistir. Diger taraftan verimi artirmaya yonelik olarak ise, (1) dosya aktarimlarinda, MD5 bazli kontroller konularak dosya aktarimi optimize edilmis ve ayni dosyalarin tekrarli aktarimi engellenmis, (2) tasarlanan onceliklendirme algoritmasi ile dugumler onceliklendirilmis, boylece dugumlerde calistirilacak uygulama alt gorevlerine oncelik degeri atama imkani verilmis, (3) dugumlerin homojen kullanimi saglanarak yaslanma oranlari dengelenmis, boylece sistemin ilk bozulma ani otelenmistir. Sistemin sinanmasi iki farkli bilgisayar ve dort farkli atanmis bilgisayar kullanarak gerceklestirilmistir. Sinanan atanmis bilgisayarlar sirasiyla, Raspberry PI Model B, Raspberry PI 2 Model B, BeagleBone Black M...

Gunumuzde akilli kartlar siklikla bilgi teknolojileri alaninda kullanilmaya basladigindan guvenli... more Gunumuzde akilli kartlar siklikla bilgi teknolojileri alaninda kullanilmaya basladigindan guvenlik onemli bir konu haline gelmistir. Bu yazida ozellikle akilli kartlarin guvenligine yonelik saldirilar anlatilmistir. Akilli kart gelistirilmesi ile ilgili calismalar sirasinda dikkat edilmesi gereken guvenlik konulari uzerinde durulmustur. Bu makale guvenli akilli kart isletim sistemi gereklerini ve akilli kart donanim ve isletim sisteminin saglamasi gereken ortak olcut guvence seviyesini belirtir. Ayrica bu makalede “dinamik guvenlik islevi” olarak adlandirilan yeni guvenlik yaklasimi ve kullanimi da anlatilmistir. : Abstract Today, as the smart cards are frequently used in Information Technologies, smart card security has become an important issue. In this paper, security attacks that are specific to smart cards are described. Security considerations that need to be taken into account when developing a secure smart card are studied. The paper indicates that the requirements of the se...

Soru yanitlama dizgeleri uzerinde surdurulen calismalar ya bilgi tabanli ya da genel agdaki veril... more Soru yanitlama dizgeleri uzerinde surdurulen calismalar ya bilgi tabanli ya da genel agdaki verileri kullanarak calismaktadir. Bu makalede tanitilan calisma bilgi tabanlidir ve Turkceye ozgudur. Ural dil ailesinin uyesi olan Turkce sondan eklemeli bir dildir ve dil bilgisi kurallari bukunlu Hint-Avrupa dillerinin dil bilgisi kurallarindan farklidir. Bu nedenle, Turkce yanitlama dizgesi Turkcenin dil bilgisi kurallarina uygun olarak hazirlanmistir. Dizge sirasiyla su asamalardan olusmaktadir: On isleme (tumcelere ayirma, sozcuklere ayirma, hecelere ayirma, ses bilimi acisindan sinama (unlu, unsuz uyumu, unlu dusmesi, bas ve sondaki harf kurallari, galatlar ve yabanci sozcukler); Bicim bilimsel cozumleme; Sozcuklerin siniflarinin belirlenmesi; Soz obeklerinin bulunmasi; Sozcuklerin rollerinin belirlenmesi; Tumcenin anlamlandirilmasi; Uygun yanitin uretilmesi. Uretilen her yanitin dogrulugu soru soran kisiye onaylatilmaktadir. Sorulan sorular ve bunlara iliskin yanitlar bir kosut derle...

Morphological units carry vast amount of semantic information for languages with rich inflectiona... more Morphological units carry vast amount of semantic information for languages with rich inflectional and derivational morphology. In this paper we show how morphosemantic information available for morphologically rich languages can be used to reduce manual effort in creating semantic resources like PropBank and VerbNet; to increase performance of word sense disambiguation, semantic role labeling and related tasks. We test the consistency of these features in a pilot study for Turkish and show that; 1) Case markers are related with semantic roles and 2) Morphemes that change the valency of the verb follow a predictable pattern.

Theory and applications of natural language processing, 2018

We present an approach to MT between Turkic languages and present results from an implementation ... more We present an approach to MT between Turkic languages and present results from an implementation of a MT system from Turkmen to Turkish. Our approach relies on ambiguous lexical and morphological transfer augmented with target side rule-based repairs and rescoring with statistical language models.

2018 3rd International Conference on Computer Science and Engineering (UBMK), Sep 1, 2018

Law enforcement authorities continually confront new and vexatious frauds involving computer viru... more Law enforcement authorities continually confront new and vexatious frauds involving computer virus attacks or credit card frauds committed against financial institutions and banking card companies. Such phenomenon used to commit online banking fraudulent transactions are occurring continuously and lead to a loss of gargantuan amount of money. Information security specialists and associations usually refer to the process of selecting practitioners as safeguards of fraud assessment, fraud analysis, or fraud detection. Designating such safeguards based on fraud calculations can be a exasperating and high-price process over the past few decades. This paper aims to understand how deep learning (DL) models can be benevolent in detecting fraudulent transactions with high accuracy. Dataset is extracted from one month of genuine financial logs of a mobile money service company in Africa, containing about more than six million transactions. The most satisfactory machine learning techniques as ensemble of decision tree (EDT), and deep learning techniques as stacked auto-encoders (SAE) and Restricted Boltzmann Machines (RBM) classifiers are applied on the preprocessed data. The performance of generated classifier models is evaluated based on accuracy, sensitivity, specificity, precision, confusion matrix and ROC values. The results of optimal accuracy are 90.49%, 80.52% and 91.53% respectively. The comparative results manifest that restricted Boltzmann machine performs superior than the other techniques.

2017 International Conference on Computer Science and Engineering (UBMK), Oct 1, 2017

Fraud detection is an enduring topic that pose a threat to banking, insurance, financial sectors ... more Fraud detection is an enduring topic that pose a threat to banking, insurance, financial sectors and information security systems such as intrusion detection systems (IDS), etc. Data mining and machine learning techniques help to anticipate and quickly detect fraud and take immediate action to minimize costs. This paper starts with the definition of intrusion detection system and its types, focuses on the implementation of a set of well-known machine learning classification algorithms (Decision Trees, Naive Bayes and Artificial Neural Networks), which can reduce the existing disadvantages of the intrusion detection systems. Experimental results on NSL-KDD dataset infer that our ANN-MLP method (Multilayer Perceptron) yields average better performance by calculating “confusion matrix” that in turn helps us to calculate performance measure such as, “Detection Rate Accuracy”, “precision” and “recall”.

Springer eBooks, 2006

This paper describes the implementation of a two-level morphological analyzer for the Turkmen Lan... more This paper describes the implementation of a two-level morphological analyzer for the Turkmen Language. Like all Turkic languages, the Turkmen Language is an agglutinative language that has productive inflectional and derivational suffixes. In this work, we implemented a finite-state two-level morphological analyzer for Turkmen Language by using Xerox Finite State Tools.

2017 International Conference on Computer Science and Engineering (UBMK), Oct 1, 2017

An intrusion detection system (IDS) monitors the network traffic looking for suspicious or malici... more An intrusion detection system (IDS) monitors the network traffic looking for suspicious or malicious activities or policy violations, which could represent an attack or unauthorized access. Traditional systems were designed to detect known attacks but cannot identify unknown threats. They most commonly detect known threats based on predefined rules or behavioral analysis through baselining the network. Expert attackers can bypass these techniques, therefore the need for more intelligent intrusion detection is increasing by the day. Intelligent IDS collects and analyzes information from different areas and deals with large amount of data, which contains various redundant and irrelevant features and results in increased overfitting, processing (training) time and low detection rate. For that reason, feature selection should be applied as a preprocessing step to improve accuracy and system performance while shrinking large data sets. In this paper, filter method which is a method used in feature selection to identify important features and to eliminate the less effective features has been developed for three different class structures on the NSL KDD dataset. After this process, classification models were created and performance measures were used to evaluate the models.

Language Resources and Evaluation, May 4, 2017

In this work, we report large-scale semantic role annotation of arguments in the Turkish dependen... more In this work, we report large-scale semantic role annotation of arguments in the Turkish dependency treebank, and present the first comprehensive Turkish semantic role labeling (SRL) resource: Turkish Proposition Bank (PropBank). We present our annotation workflow that harnesses crowd intelligence, and discuss the procedures for ensuring annotation consistency and quality control. Our discussion focuses on syntactic variations in realization of predicate-argument structures, and the large lexicon problem caused by complex derivational morphology. We describe our approach that exploits framesets of root verbs to abstract away from syntax and increase self-consistency of the Turkish PropBank. The issues that arise in the annotation of verbs derived via valency changing morphemes, verbal nominals, and nominal verbs are explored, and evaluation results for inter-annotator agreement are provided. Furthermore, semantic layer described here is aligned with universal dependency (UD) compliant treebank and released to enable more researchers to work on the problem. Finally, we use PropBank to establish a baseline score of 79.10 F1 for Turkish SRL using the mate-tool (an open-source SRL tool based on supervised machine learning) enhanced with basic morphological features. Turkish PropBank and the extended SRL system are made publicly available.

2022 7th International Conference on Computer Science and Engineering (UBMK), Sep 14, 2022