Probabilistic Reasoning

Probabilistic reasoning focuses on developing computational methods to represent and manipulate uncertainty, aiming to build systems that can reason and make decisions under conditions of incomplete information. Current research emphasizes integrating probabilistic reasoning with neural networks (neurosymbolic AI) and large language models, often employing techniques like probabilistic logical rules and sampling-based gradient estimation to improve model accuracy, explainability, and scalability. This work is significant because it addresses limitations in existing AI systems, paving the way for more robust, reliable, and explainable AI applications across diverse fields, including natural language processing, sequential recommendation, and object detection.