Position:  Postdoctoral Researcher

Current Institution:  Microsoft Research

Abstract:  Deep Learning for Language Intelligence

My research interests are mainly in the areas of machine learning, focusing on Deep Learning, and natural language processing (NLP), focusing on (visual) language intelligence. More broadly, my research goal is targeted at research domains in which deep but flexible meaning representation, strong inference, and learnable knowledge have a large payoff: for example, language intelligence robotics.  A sample, long-range application would be a personal assistant capable of interacting with many application systems to read from and affect the world.  I have conducted research at Microsoft Research AI (MSR AI), centered around two themes:  1) neural symbolic approach for natural language processing, and 2) deep learning for visual-language intelligence. On the neural symbolic approach for NLP, we developed a new network architecture:  the Tensor Product Generation Network (TPGN) for NLP, based on the general technique of Tensor Product Representations (TPRs) for encoding and processing symbol structures in distributed neural networks. On deep learning for vision-language intelligence, we proposed a hierarchically structured reinforcement learning approach to addressing the challenges of planning for generating coherent multi-sentence stories for the visual storytelling task. Finally, beyond pushing forward state-of-the-art research, I am a strong proponent of efficient and reproducible research. I also helped to ship the proposed techniques to Microsoft products and create real-world impact. We have been developing two bots, Caption Bot and Drawing Bot, in collaboration with Microsoft product team.

Bio:
Qiuyuan Huang is a researcher in the Deep Learning group at Microsoft Research AI, having joined the group as a postdoctoral researcher in 2017. She received her bachelor’s degree in 2011 from the Department of Computer Science University of Science and Technology of China (USTC) and PhD degree in 2017 from the Department of Electrical and computer engineering University of Florida. Her research interests are in the areas of deep learning and natural language processing, ranging from deep neuro-symbolic general intelligence, reinforcement learning, generative adversarial networks, visual-language intelligence and networking. She is committed to the process of bridging the gap between theory and practice and of bringing theoretical results to practical implementations. She also loves writing actual code, building real systems from scratch and making them work simply, efficiently, elegantly, and beautifully.