Detecting Cybersecurity Events from Noisy Short Text

Yagcioglu, Semih; Seyfioğlu, Mehmet Saygın; Citamak, Begum; Bardak, Batuhan; Guldamlasioglu, Seren; Yuksel, Azmi; Tatli, Emi̇n İslam

doi:10.48550/arxiv.1904.05054

Cited by 1 publication

(8 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Comparing the obtained recall, precision, and F 1 measure with preceding research [8], [34], we can approve that despite gaining not very high coefficients values, our results are comparable for the target language (Kazakh) and sometimes better for the source language (Russian).…”

Section: Evaluation Of the Resultssupporting

confidence: 79%

“…Using convolutional neural networks (CNN) and recurrent neural network models in EE in the last few years is illustrative in this regard. For example, Yagcioglu et al [8] employed CNN and a long short-term memory (LSTM) recurrent neural network to detect cyber security events from a noisy short text. The graph neural networks (GNN) use multiple neurons operating on a graph structure to enable deep learning in non-Euclidean spaces.…”

Section: A Methods Of Event Extraction From Textsmentioning

confidence: 99%

“…Following the previous studies [8], [9], [33], [34], we determine and extract three types and seven subtypes of CRE from the corpus of news articles relevant to police and criminal activities. An additional restriction is the use of a bilingual parallel corpus that includes aligned sentences in two lowresourced and under-annotated languages.…”

Section: Methodsmentioning

confidence: 99%

“…Their extraction system was based on grammatical and heuristic rules and data from Malaysian National News Agency (BERNAMA). In the paper by Yagcioglu et al [8], cybersecurity events detection was considered. Authors focused on such cybersecurity events as zero-day exploits, ransomware, data leaks, security breaches, vulnerabilities, etc.…”

Section: B Text Mining Applications In Crimementioning

confidence: 99%

“…Nevertheless, in order to prevent and investigate crimes, it is still necessary to use more efficient methods to analyze the criminal content in computer-mediated communication (CMC). One of the most challenging and efficient solutions for the identification and prediction of crime actions is crime-related event extraction (CREE) [8], [9]. Opportunities to obtain structured information about an occurred or prospective crime, accumulate this information, and then model a crime pattern, are the cause of the CREE task effectiveness.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A Parallel Corpus-Based Approach to the Crime Event Extraction for Low-Resource Languages

Khairova

Mamyrbayev

Rizun³

et al. 2023

IEEE Access

View full text Add to dashboard Cite

These days, a lot of crime-related events take place all over the world. Most of them are reported in news portals and social media. Crime-related event extraction from the published texts can allow monitoring, analysis, and comparison of police or criminal activities in different countries or regions. Existing approaches to event extraction mainly suggest processing texts in English, French, Chinese, and some other resource-rich and well-annotated languages. This paper presents a parallel corpus-based approach that follows a closed-domain event extraction methodology to event extraction from web news articles in low-resource languages. To identify the event, its arguments, and the arguments' roles in the sourcelanguage part of the corpus we utilize an enhanced pattern-based method that involves the multilingual synonyms dictionary with knowledge about crime-related concepts and logic-linguistic equations. The event extraction from the target-language part of the corpus uses a cross-lingual crime-related event extraction transfer technique that is based on supplementary knowledge about the semantic similarity patterns of the considered pair of languages. The presented approach does not require a preliminarily annotated corpus for training making it more attractive to low-resource languages and allows extracting TRANSFER, CRIME, and POLICE types of events and their seven subtypes from various topics of news articles simultaneously. Implementation of our approach for the Russian-Kazakh parallel corpus of news portals articles allowed obtaining the F1-measure of crime-related event extraction of over 82% for the source language and 63% for the target language.

show abstract

Section: Evaluation Of the Resultssupporting

confidence: 79%

Section: A Methods Of Event Extraction From Textsmentioning

confidence: 99%