Search

Syllabification may be considered trivial for humans, but it can prove to be a challenging task in terms of automated text analysis. In this study, we explore three approaches to syllabify words in Romanian using state-of-the-art deep learning architectures in sequence prediction, namely BiLSTM, CNN, and transformer. In contrast to previous approaches, our models take into account the part of speech of the word, which in return can weigh heavily in situations where words have the same written form, but different syllabification. Our best model obtains an accuracy of approximately 98% using a conditional random field on top of the BiLSTM architecture, surpassing all previous state-of-the-art models. Our model represents a building block for multiple smart learning ecosystems, ranging from better hyphenation software for text evaluation, to text-to-speech and speech-to-text frameworks employed in intelligent houses or personal assistants.

Exploring new or emerging research domains or subdomains can become overwhelming due to the magnitude of available resources and the high speed at which articles are published. As such, a tool that curates the information and underlines central entities, both authors and articles from a given research context, is highly desirable. Starting from the articles of the International Conference of Learning Analytics & Knowledge (LAK) in its first decade, this paper proposes a novel method grounded in Cohesion Network Analysis (CNA) to analyze subcommunities of authors based on the semantic similarities between authors and papers, and estimate their global impact. Paper abstracts are represented as embeddings using a fine-tuned SciBERT language model, alongside a custom trained LSA model. The extrapolation between the local LAK community to a worldwide importance was also underlined by the comparison between the rankings obtained from our method and statistics from ResearchGate. The accuracies for binary classifications in terms of high/low impact predictions were around 70% for authors, and around 80% for articles. Our method can guide researchers by providing valuable information on the interactions between the members of a knowledge community and by highlighting central local authors who may potentially have a high global impact.

The use of computational systems for enhancing collaborative learning is widely spread. While interest in the application of Computer-Supported Collaborative Learning (CSCL) systems in cybersecurity exists, endeavors in this sense are difficult to compare due to different choices of methods and tools. In this sense, we advance a typology of CSCL systems in cybersecurity and introduce an automated system, Yggdrasil, which identifies emerging vulnerability threats from tweets containing references to explicit news articles. Such a system can be used to enhance collaborative learning processes of experts and non-experts in cybersecurity regarding the early detection of vulnerabilities, by automatically extracting information from online communities. For this purpose, we built a corpus of 650 annotated tweets that contain links to cybersecurity related articles, posted by key profiles in the cybersecurity landscape. Using insights from state-of-the-art approaches in early detection of vulnerabilities from tweets and sentiment analysis from Twitter, we conducted two experiments for the development of Yggdrasil. Our results argue that the use of the BERT language model is an adequate approach for identifying relevant posts or news articles, having an accuracy over 91% in various configurations. Furthermore, we also illustrate the use of transfer learning as a viable alternative for the early detection of cybernetic vulnerabilities.

The process of identifying relevant scientific papers and authors is becoming more and more tedious both to neophytes, as well as to experts searching for state-of-the-art approaches, due to the high-speed development of recent emerging domains and the exponential increase of the available publications. This paper presents an overarching architecture grounded in Cohesion Network Analysis (CNA) and provides a detailed showcase on the papers published at the Learning Analytics & Knowledge (LAK) conference between 2011 and 2020. We applied our method to LAK because it is the most representative sub-community from the Learning Analytics domain. On the other hand, CNA Sociograms might be applied to journals, academic departments, business groups, etc., having the potential to aid in understanding the past and current foci of a community and its directions. We propose an end to end method, from crawling the ACM Digital Library to insights presentation, using our own Visual 2-Mode CNA Graph and visualizations in Kibana, customizable and interactive, that facilitates individuals' interactions with a new domain. Our method brings valuable insights about the semantic relatedness between authors and other types of associations presented as links within the resulting 2-mode cohesion graph. Moreover, we introduce multiple statistics about top ranked authors, institutions, as well as the most central domain articles. In contrast to previous experiments, we introduce a comprehensive use case, an enhanced processing pipeline with new heuristics, as well as new CNA sociogram visualizations. This semantic-based approach facilitates inferring key authors' and publications' interrelations and trending topics over time.

Reading comprehension is key to knowledge acquisition and to reinforcing memory for previous information. While reading, a mental representation is constructed in the reader’s mind. The mental model comprises the words in the text, the relations between the words, and inferences linking to concepts in prior knowledge. The automated model of comprehension (AMoC) simulates the construction of readers’ mental representations of text by building syntactic and semantic relations between words, coupled with inferences of related concepts that rely on various automated semantic models. This paper introduces the second version of AMoC that builds upon the initial model with a revised processing pipeline in Python leveraging state-of-the-art NLP models, additional heuristics for improved representations, as well as a new radiant graph visualization of the comprehension model.

Dialogism is a philosophical theory centered on the idea that life involves a dialogue among multiple voices in a continuous exchange and interaction. Considering human language, different ideas or points of view take the form of voices, which spread throughout any discourse and influence it. From a computational point of view, voices can be operationlized as semantic chains that contain related words. This study introduces and evaluates a novel method of identifying semantic chains using BERT, a state-of-the-art language model for computational linguistics. The resulting model generalizes to multiple relations including repetitions, semantically related concepts from WordNet (i.e., synonyms, hypernyms, hyponyms, and siblings), as well as pronominal resolutions. By combining the attention scores between words, word pairs are merged into connected components that denote emerging voices from the discourse. The introduced visualization argues for a more dense capturing of inner semantic links between words and even compound words in contrast to classical methods of building lexical chains.

The drawbacks of traditional methods of cybernetic vulnerability detection relate to the required time to identify new threats, to register them in the Common Vulnerabilities and Exposures (CVE) records, and to score them with the Common Vulnerabilities Scoring System (CVSS). These problems can be mitigated by early vulnerability detection systems relying on social media and open-source data. This paper presents a model that aims to identify emerging cybernetic vulnerabilities in cybersecurity news articles, as part of a system for automatic detection of early cybernetic threats using Open Source Intelligence (OSINT). Three machine learning models were trained on a novel dataset of 1000 labeled news articles to create a strong baseline for classifying cybersecurity articles as relevant (i.e., introducing new security threats), or irrelevant: Support Vector Machines, a Multinomial Naive Bayes classifier, and a finetuned BERT model. The BERT model obtained the best performance with a mean accuracy of 88.45% on the test dataset. Our experiments support the conclusion that Natural Language Processing (NLP) models are an appropriate choice for early vulnerability detection systems in order to extract relevant information from cybersecurity news articles.

Execution of complex tasks can be optimized by using parallel computing integrated within specialized frameworks. Thus, the AKKA framework was employed to speed up the processing time of the HUB-TECH platform to recommend semantically relevant articles, starting from a project description. Both parallel and distributed approaches are presented and a speed-up analysis between the two is performed. Distributed architectures enhance the system's performance by computing an increased number of semantic distances in less time. This usually requires more physical resources than parallel processing which is also capable of efficiently exploiting single nodes. The hardware specifications of the machine are relevant when trying to determine the optimal number of actors that can run on a machine in order to produce the maximal throughput. Our results show a 22% decrease in execution time while comparing serial and parallel approaches, a drastic 86% decrease when distributed methods are employed in contrast to serial executions, as well as an 82% improvement when comparing distributed and parallel experiments.

Online searching is one of the most frequently performed actions and search engines need to provide relevant results, while maintaining scalability. In this paper we introduce a novel approach grounded in Cohesion Network Analysis in the form of a semantic search engine incorporated in our Hub-Tech platform. Our aim is to help researchers and people unfamiliar with a domain find meaningful articles online, relevant for their project scope. In addition, we integrate state-of-the-art technologies to ensure scalability and low response time, namely SOLR – for data storage and full-text search functionalities – and Akka – for parallel and distributed processing. Preliminary validations denote promising search results, the software being capable to suggest articles in approximately the same way as humans consider them most appropriate – 75% are close results and top 20% are identical to user recommendations. Moreover, Hub-Tech recommended more suitable articles than Google Scholar for our specific task of searching for articles related to a detailed description given as input query (50 + words).

Nowadays blogging has become one of the most common and frequently employed methods for creating communities around specific topics of interest. This phenomenon is due to the existence of numerous online platforms that enable the creation with ease of blogs even for non-technical users. However, finding relevant blogs is a difficult and time-consuming process. This paper focuses on predicting whether a post has an impact or not based only on its writing style and automatically inferred textual complexity indices. We define the impact of a post as the traction it receives from other users in terms of number of received comments. Thus, several blogs were crawled on which a set of textual complexity indices available from the ReaderBench framework were computed. By splitting the posts in two categories, we were able to predict with an accuracy of 69.7% whether a post will receive comments or not.