A Multi-Modal AI Decision Framework for Video Content Moderation Using FNSS, Deep Learning, and NLP
Year: 2026
Type: conference-paper
Abstract: Nowadays, kids are highly engaged with social media platforms, and there is always a chance that they may get to see inappropriate videos that are not suitable for their age group. The main aim of this research is to provide a complete solution to the issue by filtering the explicit video content using the concept of computer vision, natural language processing (NLP), and uncertainty modeling using Fermatean Neutrosophic Soft Sets (FNSS). The significant advantage of this system is that it provides a feedback loop using the concept of machine learning, which means that the system can be improved by incorporating feedback from the parents or moderators to make the best decisions by changing the thresholds of the decisions based on real-time data. The article describes each step of the system, from ingesting the data to the adaptive feedback loop with the help of mathematical concepts and the execution time of each step. In order to test the system and confirm that the system is actually effective in the real world, the Proof-of-Concept (POC) test was conducted using the YouTube video by Parinaz Shajareh. The system is effective in the real world with the inclusion of ethical factors, and the execution time is highly optimized.
Institution: Jadara University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7166017175