Ziying Guo

About

Ziying Guo is a researcher at the forefront of vision-based robotic manipulation, with a focus on bridging perception and action in unstructured environments. Her work centers on two key challenges: how robots can learn meaningful object representations, and how they can acquire manipulation skills through reinforcement learning. In her highly cited 2020 paper, Guo introduces the concept of learning an “affordance space” from visual input, arguing that robots, like humans, should infer what actions an object enables—such as grasping or pushing—rather than focusing on superficial features like texture or lighting. This affordance-driven approach has garnered 14 citations and offers a principled framework for generalizable manipulation. Earlier, in 2018, she explored domain centralization and cross-modal reinforcement learning to improve robotic performance in agricultural tasks like picking and sorting, demonstrating how deep reinforcement learning can be adapted to real-world, visually complex settings. Though her citation counts are still growing, Guo’s work is notable for its conceptual clarity and practical relevance, positioning her as an emerging voice in embodied AI and autonomous robotics.

Research Focus

Key Achievements

Top Papers

1. 1
   Learning Affordance Space in Physical World for Vision-based Robotic Object Manipulation

   14 citations · 2020
2. 2
   Domain centralization and cross-modal reinforcement learning for vision-based robotic manipulation

   2 citations · 2018

Key Collaborators

Contact & Links