Enhancing Zero-shot Commonsense Reasoning by Integrating Visual Knowledge via Machine Imagination

Researchers propose 'Imagine,' a new zero-shot commonsense reasoning framework that enhances Pre-trained Language Models by integrating machine-generated visual signals into the reasoning pipeline. The approach demonstrates superior performance over existing zero-shot methods and even advanced large language models by addressing human reporting biases through machine imagination.

• →The Imagine framework supplements textual inputs with visual signals from machine-generated images to improve AI reasoning.
• →The approach addresses human reporting biases that limit current Pre-trained Language Models in commonsense reasoning tasks.
• →Synthetic datasets were constructed to emulate visual question-answering scenarios for effective visual context utilization.
• →Comprehensive evaluations show Imagine outperforms existing zero-shot approaches and advanced large language models.
• →Machine imagination demonstrates potential to significantly enhance generalization abilities in AI reasoning models.