Filter Bubble

Filter bubbles describe the phenomenon where personalized online systems, such as recommender systems and large language models, limit users' exposure to diverse perspectives, reinforcing existing biases. Current research focuses on understanding the formation of these bubbles across various platforms (e.g., short-video, music streaming) and developing algorithms, like weighted hypergraph embedding, to improve recommendation diversity without sacrificing accuracy. This research is crucial for mitigating the negative impacts of filter bubbles on user satisfaction, polarization, and the spread of misinformation, informing the design of more equitable and inclusive online experiences.