Description:
In the current state of the rapid growth of information resources and the huge number of requests submitted by users to existing information retrieval systems; recently, Question Answering systems have attracted more attention to meet information needs providing users with more precise and focused retrieval units. As one of the most challenging and important processes of such systems is to retrieve the best related text excerpts with regard to the questions, we propose a novel approach to exploit not only the syntax of the natural language of the questions and texts, but also the semantics relayed beneath them via a semantic question rewriting and passage retrieval task. The semantic structure used to address the surface mismatch of the semantically related passages and queries is FrameNet which is a lexical resource for English constituted based on frame semantics. We have run our proposed approach on a subset of the TREC 2004 factoid questions to retrieve passages containing correct answers from the AQUAINT collection and we have obtained promising results.
Description:
In the current state of the rapid growth of information resources and the huge number of requests submitted by users to existing information retrieval systems; recently, Question Answering systems have attracted more attention to meet information needs providing users with more precise and focused retrieval units. As one of the most challenging and important processes of such systems is to retrieve the best related text excerpts with regard to the questions, we propose a novel approach to exploit not only the syntax of the natural language of the questions and texts, but also the semantics relayed beneath them via a semantic question rewriting and passage retrieval task. The semantic structure used to address the surface mismatch of the semantically related passages and queries is FrameNet which is a lexical resource for English constituted based on frame semantics. We have run our proposed approach on a subset of the TREC 2004 factoid questions to retrieve passages containing correct answers from the AQUAINT collection and we have obtained promising results.
Description:
In this technology-based era, network-based systems are facing new cyber-attacks on daily bases. Traditional cybersecurity approaches are based on old threat-knowledge databases and need to be updated on a daily basis to stand against new generation of cyber-threats and protect underlying network-based systems. Along with updating threat-knowledge databases, there is a need for proper management and processing of data generated by sensitive real-time applications. In recent years, various computing platforms based on representation learning algorithms have emerged as a useful resource to manage and exploit the generated data to extract meaningful information. If these platforms are properly utilized, then strong cybersecurity systems can be developed to protect the underlying network-based systems and support sensitive real-time applications. In this survey, we highlight various cyber-threats, real-life examples, and initiatives taken by various international organizations. We discuss various computing platforms based on representation learning algorithms to process and analyze the generated data. We highlight various popular datasets introduced by well-known global organizations that can be used to train the representation learning algorithms to predict and detect threats. We also provide an in-depth analysis of research efforts based on representation learning algorithms made in recent years to protect the underlying network-based systems against current cyber-threats. Finally, we highlight various limitations and challenges in these efforts and available datasets that need to be considered when using them to build cybersecurity systems.