moviQuest-MAS: Engineering Intelligence into Ubiquitous Social Business

TL;DR

moviQuest-MAS is a specialized Multi-Agent System (MAS) platform designed to transform social networking into a "serious" business and academic tool. By combining a web/desktop architecture with "Trainedphones"—feature phones equipped with lightweight expert systems—it enables secure, intelligent interaction even in low-bandwidth environments using encrypted SMS and fuzzy logic.

Context: Beyond the "Gossip" of Social Media

While social media has revolutionized communication, the authors argue that mainstream platforms (like Facebook) are often unsuited for professional or academic environments. The noise-to-signal ratio is too high, and "serious" conversations often get lost in casual gossip. Furthermore, there is a technical divide: sophisticated AI features usually require high-end smartphones and constant data connectivity.

The moviQuest-MAS project seeks to bridge this gap by introducing Ubiquitous Artificial Intelligence, focusing on affinity groups where interaction is strictly governed by shared professional or community interests.

Methodology: The Multi-Agent Ecosystem

The core of the system is governed by the Agent-Group-Role (AGR) model, which defines how human and software agents interact within specific contexts.

1. The Software Ecosystem

The platform is not just an app but a three-tier ecosystem:

• Control Center: A desktop application for system managers to handle sensitive data away from web-based risks.
• Social Network Engine: The web portal for standard microblogging and group management.
• Mobile Client: A unique application that uses an specialized ontology to communicate via encrypted SMS.

2. The Multi-Agent Architecture

The system logic is expressed through the tuple ${\Xi }_{\alpha }=(\mathbf{A},\mathbf{G},\mathbf{R},{\mathcal{K}}_{\alpha })$. This architecture allows for seamless scaling—new functionalities (like "Spontaneous Evaluation" or "Market Monitoring") are simply added as new software agent roles.

Figure 1: The Multi-Agent model showing the interaction between human agents, software robots, and the knowledge base.

Technical Innovation: Trainedphones and MQDM

A standout feature of this research is the concept of the Trainedphone. Unlike "smartphones" that rely on superior hardware, a Trainedphone is a standard device "taught" to be smart via a Mobile Smart Agent. This is a light expert system that uses a tree-grammar ontology to:

• Compress complex surveys into 1-3 SMS messages.
• Manage intelligent interview sequences locally on the device to minimize user effort.

Furthermore, the paper introduces MQDM (moviQuest Decision Making). This is a fuzzy aggregation algorithm that allows organizations to conduct sentiment analysis and group decision-making. It doesn't just count votes; it weighs them based on:

• The user's record of accuracy in past decisions.
• The user's self-declared expertise on a specific topic.
• The "intention" embedded in the response.

Practical Implementation: QuoSity

The authors validated their framework through QuoSity, a social network for academic environments. Unlike Facebook, which struggles to separate a student's private life from their academic responsibilities, QuoSity uses the MAS affinity groups to naturally silo conversations.

Figure 2: The QuoSity platform implementing the moviQuest-MAS features in a pedagogical setting.

Critical Insight & Conclusion

The significance of moviQuest-MAS lies in its inclusive approach to AI. By optimized encoding through a specific ontology, the authors prove that sophisticated multi-agent coordination does not require 5G or the latest iPhone. It can be built on the "ubiquitous" infrastructure of SMS and basic mobile computing.

Takeaways:

• Context is King: Social networks are more valuable when restricted by affinity and professional roles.
• Software-Defined Intelligence: Expert systems can make "dumb" phones smart enough for enterprise-grade data collection.
• Fuzzy Governance: Group decision-making in the digital age requires more than just majority rule—it requires weight-based, intelligent aggregation.

While the reliance on SMS might seem dated in an era of ubiquitous 4G/5G, the underlying logic of bandwidth-efficient MAS communication remains highly relevant for satellite IoT and secure, air-gapped enterprise environments.