An Efficient SuperHyperSoft Framework for Evaluating LLMs-based  Secure Blockchain Platforms

Mona Mohamed; Alaa Elmor; Florentin Smarandache; Ahmed A. Metwaly

Authors

Mona Mohamed Higher Technological Institute, 10th of Ramadan City 44629, Egypt
Alaa Elmor Zagazig University, 44519 Zagazig, Egypt
Florentin Smarandache University of New Mexico, 705 Gurley Ave., Gallup, NM 87301, USA
Ahmed A. Metwaly Zagazig University, 44519 Zagazig, Egypt

Keywords:

Large Language Models (LLMs); SuperHyperSoft; Blockchain; Cybersecurity; Multi Criteria Decision Making (MCDM); Single Value Neutrosophic (SVN)

Abstract

In the age of modern technology, the use of the internet has become imperative.
However, this widespread access presents a double-edged sword and opens doors for hackers and
scammers to exploit vulnerabilities and engage in illegal activities. Accordingly, scholars and
stakeholders are attempting to solve this matter. Large Language Models (LLMs) have provided
highly effective methodologies and solutions in various cybersecurity sectors. Hence, we exhibited
the efficacy of LLMs in several information and communication technologies (ICT) such as the
Internet of Things (IoT), cloud computing, blockchain technology (BCT)…etc which are attacked
and threatened. Accordingly, the objective of our study is to illustrate how LLMs are supporting
ICT, especially BC to be secure against attacks. Another study’s objective is to aid the stakeholders
and enterprises that seek resilience and sustainability by recommending the most secure BC
platform to adopt in critical sectors. Wherein, LLMs support BC in many directions as developing
secure smart contracts and scanning the smart contract to protect it from any subversive acts by
identifying anomalous activities. Hence, we suggested a soft opting model to rank the alternatives
of BC platforms and recommend optimal BC. Also, the process of constructing this model requires
leveraging several techniques. We applied for the first time SuperHyperSoft (SHS) as an extension
of Hypersoft to treat various attributes and sub-attributes for BC based on LLMs. Multi-criteria
decision-making (MCDM)techniques are utilized for their ability to treat conflicting sub-attributes.
Hence, entropy and multi-objective optimization based on simple ratio analysis (MOOSRA) are
utilized as techniques of MCDM. These techniques are working under the authority of the Single
Value Neutrosophic (SVN) technique to support MCDM techniques in ambiguous situations.