Dr. Siqi Bao is a senior R&D engineer at Baidu NLP, where she leads the foundational research of PLATO. She obtained her B.E. from Zhejiang University in 2012 and her Ph.D. from Hong Kong University of Science and Technology in 2017. Her research interests center around dialogue systems, pre-trained models, natural language processing, and medical image analysis. Recently, she has focused on aligning LLMs with human values, while also driving the integration of LLMs into vital products.

In 2017, Siqi joined Baidu NLP, embarking on a dedicated journey in dialogue systems. As the team leader, she and her team achieved multiple championships in DSTC9 and DSTC10, showcasing the capabilities of PLATO’s pre-trained dialogue model. Their contributions extended beyond competition, as they played a pivotal role in successfully integrating PLATO into key Baidu products. In 2021, Siqi and her team were honored with the “百度最高奖” in recognition of their outstanding work. In 2022, Siqi and Baidu’s team received the esteemed “吴文俊人工智能科技进步奖 特等奖”, highlighting their breakthroughs in general dialogue systems.

We are hiring (both interns and employees)! If you are passionate about foundation models and alignment research, please send us an email with your resume.

Research Highlights

• [PLATO] is a series of pre-trained dialogue models in English and Chinese.
  • PLATO proposes to model the one-to-many relationship in open-domain conversation with latent variables.
  • PLATO-2 builds the large-scale pre-trained dialogue models effectively via curriculum learning.
  • PLATO-XL releases the first 11B parameter English and Chinese pre-trained dialogue model.
  • PLATO-K introduces the pre-trained dialogue model with comprehensive internal & external knowledge enhancement.