UROP Project

The Logic of the Feed: Mathematical Models of Recommender Systems in Digital Media Platforms

Economics, Game Theory, Recommender Systems, Human-AI Interaction

Research Mentor: Prof. Marcos Müller Vasconcelos,
Department, College, Affiliation: Electrical Engineering, FAMU-FSU College of Engineering
Contact Email: m.vasconcelos@fsu.edu
Research Assistant Supervisor (if different from mentor):
Research Assistant Supervisor Email:
Faculty Collaborators:
Faculty Collaborators Email:

Looking for Research Assistants: Yes
Number of Research Assistants: 2
Relevant Majors: Economics, Statistics, Business, Electrical Engineering, Computer Science
Project Location: FAMU-FSU College of Engineering
Research Assistant Transportation Required: Yes
Remote or In-person: Partially Remote
Approximate Weekly Hours: 10 hours/week, Flexible schedule (Combination of business and outside of business. TBD between student and research mentor.)
Roundtable Times and Zoom Link:
Not participating in the roundtable

Project Description

Most information consumed online today flows through digital media platforms such as YouTube, TikTok, and Instagram. Because the volume of available content is overwhelming, these platforms rely on recommendation systems to filter and promote material. However, recommendations are made under uncertainty: platforms cannot directly observe a user’s private preferences. Instead, they adapt their strategies to maximize engagement, even if it means amplifying extreme or polarized content. Our recent research has modeled this interaction as a signaling game between a platform and a user: the user seeks content aligned with their preferences, while the platform seeks to maximize engagement regardless of alignment. Theoretical results show that equilibrium strategies in such games can naturally lead to the escalation of extreme content availability. This raises important questions about how algorithms, user behavior, and economic incentives jointly shape the online information ecosystem.

Objectives:

This UROP project aims to explore recommender systems by focusing on three interconnected areas:

1. Algorithmic Foundations – how recommendation strategies influence exposure to extreme versus moderate content.

2. User Behavior & Engagement – how individuals respond to content depending on alignment and private preferences.

3. Economic & Game-Theoretic Aspects – how engagement-driven incentives can lead to unintended outcomes, such as polarization and escalation of content intensity.

This is a joint project with professors from the University of Southern California (USC).


Research Tasks: 1. Simulation & Modeling: Build simplified game-theoretic or agent-based models of user–platform interactions, exploring equilibrium dynamics.

2. Data Analysis: Use small-scale datasets (e.g., Reddit, Twitter/X, or simulated feeds) to study recommendation patterns and engagement responses.

3. Theoretical Exploration: Extend the signaling game framework to test how assumptions (e.g., user heterogeneity, platform objectives) affect equilibrium outcomes.

Skills that research assistant(s) may need: Students should have:

1. Basic programming (Python, MATLAB, or R) - recommended

2. Probability/statistics background - recommended

3. Interest in digital media, algorithms, and game theory - required