Lab of Computer Networks & Mobile Systems (NeMo)

With the intensification of internet censorship measures implemented across the globe, detecting and analyzing censorship activities have become crucial for preserving Internet freedom. This study proposes leveraging Large Language Models (LLMs) such as GPT and BERT to tackle these challenges. By drawing on parallels between network reachability data and NLP corpus—particularly their scale, structure and semantic depth—we explore how LLMs can enhance censorship detection within vast datasets. This approach aims to increase both the accuracy and scalability of censorship detection, making censorship data more accessible to downstream data consumers such as policy makers, journalists, and researchers in censorship circumvention community.

In our current high-tech world, with innovations such as the internet and artificial intelligence leading to new frontiers in development, such as large language models, self-driving cars, and deepfakes, we live in unprecedented times, where computer security is more important than ever. However, one difficulty in terms of the implementation of computer security lies in the weak point of vulnerabilities due to actions of the end user not following best practices, such as unencrypted documents and reuse of passwords, leading to stolen data, and in more extreme cases, identity theft. This project aims at understanding the public perceptions of cybersecurity in terms of knowledge of and tolerance to cybersecurity and privacy, based on multiple criteria, and the development of a framework for cybersecurity education. 

While the world has been reshaped by advances in computing and communication,the most traditional institution of passing on knowledge – Schools remains unchanged with old modalities. Further, because of inequity in resources and awareness, digital divides between those who are privileged and those who struggle educationally and economically are even widened. In this project, we aim to address the challenges of Attainability and Equity in Education, particularly we would like to design a new education ecosystem – Attainable Education for All (AeDA) – that rides on digital technologies and services to redefine, transform, and foster modern educational practices in a model that college level curriculum are structured with digital economy job market demands in mind; and furthermore instead of documenting educational achievements/goals (such as certificates or degrees) on the basis of courses, the AeDA ecosystem will enable degree studies to be established on the basis of knowledge points for individuals by obtaining customized training for particular job placements. We envision the AeDA ecosystem will increase attainability, advance the benefits of peer engagements, respect the diversity of learners, and bridge the gap between job market demand and workforce development in the digital economy.