Meta has introduced TRIBE v2, an advanced artificial intelligence model designed to understand human brain behaviour by predicting neural responses based on visual, audio and text inputs. The model has been trained on large datasets and participant inputs to improve the speed and accuracy of research.
TRIBE v2 processes visual, audio and text inputs simultaneously to estimate how the human brain responds. Meta stated that the model delivers more accurate and detailed outputs compared to its previous version, supported by training on larger datasets and a greater number of participants.
Traditional brain-related experiments require fresh brain recordings for each study, a process that is time-consuming and expensive. TRIBE v2 is designed to simplify this process, enabling faster research. According to Meta, studies that previously required months may now be completed within minutes, potentially impacting neuroscience, healthcare and artificial intelligence research.
The model operates in three stages. It first converts visual, audio and text inputs into data formats for analysis. It then integrates these inputs to identify patterns in human brain activity. In the final stage, it predicts which areas of the brain are activated in response to specific inputs, distinguishing it from conventional AI models.
