Why am I seeing this? Towards recognizing social media recommender systems with missing recommendations

TL;DR

The paper addresses identifying the hidden recommender system shaping user interactions on social platforms under data-access constraints. It introduces a Graph Neural Network (GNN) framework that learns a Recommender-Neutral User (RNU) model, creates Recommender Hypothesis-specific Synthetic Datasets (RHSD), and trains Recommender Hypothesis-specific User (RHU) models to compare likelihoods under different recommender hypotheses, leveraging the loss $L(\theta; D, R) = - \sum_{i=1}^N \log P_\theta(y_i | x_i, R)$. Key contributions include the RNU design, two approaches to handle hidden recommenders (marginalization and hindsight predictive modeling), and the use of Infospheres to generate diverse synthetic data, with results showing that matching recommender hypotheses generally maximize the likelihood, validating hidden-recommender inference from interactions alone. The work provides a scalable, behavior-based auditing alternative that can inform simulations and regulatory efforts to mitigate polarization and misinformation on social networks. By applying academic networks as proxies for social-media recommenders, the approach offers practical insights into how different recommender strategies shape user actions and network dynamics.

Abstract

Social media plays a crucial role in shaping society, often amplifying polarization and spreading misinformation. These effects stem from complex dynamics involving user interactions, individual traits, and recommender algorithms driving content selection. Recommender systems, which significantly shape the content users see and decisions they make, offer an opportunity for intervention and regulation. However, assessing their impact is challenging due to algorithmic opacity and limited data availability. To effectively model user decision-making, it is crucial to recognize the recommender system adopted by the platform. This work introduces a method for Automatic Recommender Recognition using Graph Neural Networks (GNNs), based solely on network structure and observed behavior. To infer the hidden recommender, we first train a Recommender Neutral User model (RNU) using a GNN and an adapted hindsight academic network recommender, aiming to reduce reliance on the actual recommender in the data. We then generate several Recommender Hypothesis-specific Synthetic Datasets (RHSD) by combining the RNU with different known recommenders, producing ground truths for testing. Finally, we train Recommender Hypothesis-specific User models (RHU) under various hypotheses and compare each candidate with the original used to generate the RHSD. Our approach enables accurate detection of hidden recommenders and their influence on user behavior. Unlike audit-based methods, it captures system behavior directly, without ad hoc experiments that often fail to reflect real platforms. This study provides insights into how recommenders shape behavior, aiding efforts to reduce polarization and misinformation.