Yaobo Liang

关于

Yaobo Liang (梁耀波) is a Researcher at Microsoft Research Asia (MSRA). He joined MSRA at March 2016. He is working on Embodied AI, AI Agents, and Cross-lingual Pre-training.

He received his Master(2016) and Bachelor(2013) degree from Beijing Institute of Technology.

Highlights:

• AI Agents
  • TaskMatrix.AI: Completing Tasks by Connecting Foundation Models with Millions of APIs (paper (opens in new tab))
  • AI Astronomer: Interpreting Multi-band Galaxy Observations (paper (opens in new tab))
  • Learning to plan with natural language (paper (opens in new tab))
• LLM/Agent Evaluations
  • AGIEval: A Human-Centric Benchmark for Evaluating Foundation Models (paper (opens in new tab))
  • PPTC benchmark: Evaluating large language models for powerpoint task completion (paper (opens in new tab))
  • PPTC-R benchmark: Evaluating the Robustness of Large Language Models for PowerPoint Task Completion (paper (opens in new tab))
• Computational Linguistics
  • Machine-Created Universal Language for Cross-lingual Transfer (paper (opens in new tab))
  • Composition is the key for emerging cross-lingual ability without parallel data (paper (opens in new tab))
  • Discovering language relation from multilingual pre-training (paper (opens in new tab))
• Dense Retrieval
  • MVR (paper (opens in new tab)) Generate multiple representation from different view.
  • CCP (paper (opens in new tab)) Generate aligned cross-lingual sentence representation without bilingual data.
• XLM-K (paper (opens in new tab)) Cross-lingual pre-training with multilingual knowledge base.
• XGLUE (paper (opens in new tab)) is a new benchmark dataset for cross-lingual understanding and generation.