Yuyi Wang

About

Yuyi Wang is a researcher at the forefront of artificial intelligence, specializing in deep reinforcement learning, multimodal learning, and autonomous navigation. Wang’s most influential contribution is pioneering the use of deep reinforcement learning to teach machines how to read 2D maps for navigation in complex 3D environments—a task that challenges even humans. This work, presented in two highly cited papers (2017, 2018), has garnered over 50 citations and addresses a critical problem in robotics and AI: enabling agents to localize and navigate using spatial representations. More recently, Wang has advanced the field of large multimodal models, developing unified algorithms that learn and generate across text, images, and video using next-token prediction—a breakthrough that extends the success of large language models into multimodal domains. This emerging work, already cited in 2026, promises to reshape how AI systems process and integrate diverse sensory information. Wang’s research sits at the intersection of spatial reasoning and multimodal intelligence, with significant implications for autonomous systems, robotics, and human-AI interaction.

Research Focus

Key Achievements

Top Papers

1. 1
   Teaching a Machine to Read Maps With Deep Reinforcement Learning

   38 citations · 2018
2. 2
   Teaching a Machine to Read Maps with Deep Reinforcement Learning

   16 citations · 2017
3. 3
   Multimodal learning with next-token prediction for large multimodal models

   3 citations · 2026

Key Collaborators

Contact & Links