Yongxin Yang

Papers

Total Citations

H-Index

About

Yongxin Yang is a leading researcher in machine learning, with a primary focus on computer vision, natural language processing, and reinforcement learning. His most influential work bridges the gap between visual perception and language generation, most notably through his 2017 paper "Actor-Critic Sequence Training for Image Captioning," which has garnered 99 citations. In this seminal contribution, Yang introduced a novel reinforcement learning framework that directly optimizes non-differentiable evaluation metrics like CIDEr and BLEU, overcoming the limitations of traditional maximum likelihood estimation. This approach significantly improved the quality and relevance of generated image descriptions, enabling AI agents to better communicate with human users about visual content. Yang's work has been instrumental in advancing the field of visual intelligence, particularly for applications in robotics and human-AI interaction where accurate scene understanding and natural language communication are critical. His research continues to influence how machines learn to describe the visual world, making him a notable figure in the intersection of vision and language.