Max Robotics — Global Service Robot Directory + US Certification

About

Weikun Guo is a robotics researcher advancing the frontier of few-shot learning and meta-learning for robotic manipulation. His work focuses on enabling robots to rapidly adapt to novel objects and environments with minimal demonstrations—a critical challenge for deploying versatile robots in unstructured settings. Guo’s most cited paper, “Few-Shot Instance Grasping of Novel Objects in Clutter” (9 citations), introduces a meta-learning framework that allows a robot to grasp a specific target object from a cluttered scene after seeing only a handful of examples. This work directly addresses the “instance grasping” problem, where the robot must not just pick any object but the correct one. He further extends these ideas in “Learning With Dual Demonstration Domains” (8 citations) and “Replayed Task-Contrastive Model-Agnostic Meta-Learning” (5 citations), developing domain-adaptive meta-learning algorithms that transfer skills across different visual environments. His 2022 papers collectively establish a systematic approach to learning from visual demonstrations, incorporating multi-level attention mechanisms to handle domain shifts. With a total of over 26 citations in just two years, Guo’s contributions are shaping how robots can learn new tasks quickly and robustly, bringing us closer to truly adaptive robotic assistants.