Skip to main content
Parents & Friends homeWPI News home
Story
2 of 25

Robotics Professor Constantinos Chamzas Awarded $175K NSF Grant to Advance Robot Learning

Constantinos Chamzas Professor Constantinos Chamzas, a faculty member in the Department of Robotics Engineering, has been awarded a prestigious $175,000 grant from the National Science Foundation (NSF) to support his research in robotic planning and manipulation. The award, part of the NSF’s highly competitive Computer and Information Science and Engineering Research Initiation Initiative (CRII), will help launch a project titled “CRII: Towards Real-World Robotic Manipulation: Learning Abstract State and Action Representations from Visual and Execution Data” which aims to revolutionize how robots learn and reason in complex, real-world environments.    Professor Chamzas’s inspiration for the project traces back to his doctoral research, where he explored how robots leverage past experiences to improve planning efficiency. “I’ve always been fascinated by how classical planning algorithms offer strong generalization in theory,” he explains, “but in practice, they require carefully designed spaces and significant manual effort to function effectively.” His curiosity led him to explore the intersection of symbolic planning and machine learning—two traditionally distinct approaches in robotics.  The core idea behind his project is deceptively simple: enable robots to reason more like humans. When we put clothes in a closet,  “We don’t consciously model every object or constraint,” Chamzas says. “We just follow an abstract plan: go to the closet, open the door, put the clothes inside.” But for a robot, that same task requires a detailed, manually encoded model. His research seeks to change that by allowing robots to learn abstract representations of tasks and actions directly from experience, rather than relying on human-specified models.    Technically, the project focuses on enabling robots to perform long-horizon manipulation tasks by learning symbolic abstractions from real-world data.  “Instead of assuming a perfect model of the world,” Professor Chamzas states, “the robot will autonomously collect and analyze its own experience to discover how to represent tasks and actions symbolically.” The result enables more adaptive and explainable robotic behavior.    The grant application process, Professor Chamzas notes, was both challenging and rewarding. “The CRII program is unique in how it supports early-career researchers,” he says. It gave him the “opportunity to distill my long-term research vision into a focused, high-impact proposal.” He credits the support of his colleagues in the Robotics Department and past CRII recipients for helping him refine his ideas and navigate the application process.     For other researchers seeking NSF funding, Professor Chamzas offers practical advice: “Start early and don’t be afraid to share your ideas with trusted peers and mentors. Treat the proposal not just as a funding opportunity, but as a chance to clarify and articulate your long-term research vision” He emphasizes the importance of grounding proposals in prior work and being open to feedback—even when it’s conflicting. Professor Chamzas says that open dialogue with colleagues is what helped him the most, and he strongly encourages open conversations.     With this NSF grant, Professor Chamzas is poised to make significant strides in the field of robotics, pushing the boundaries of how machines learn, plan, and interact with the world. The work supported by this award will contribute to broader developments in the field and provide valuable insights for the robotics community at large.  

Constantinos Chamzas

Professor Constantinos Chamzas, a faculty member in the Department of Robotics Engineering, has been awarded a prestigious $175,000 grant from the National Science Foundation (NSF) to support his research in robotic planning and manipulation. The award, part of the NSF’s highly competitive Computer and Information Science and Engineering Research Initiation Initiative (CRII), will help launch a project titled “CRII: Towards Real-World Robotic Manipulation: Learning Abstract State and Action Representations from Visual and Execution Data” which aims to revolutionize how robots learn and reason in complex, real-world environments.   

Professor Chamzas’s inspiration for the project traces back to his doctoral research, where he explored how robots leverage past experiences to improve planning efficiency. “I’ve always been fascinated by how classical planning algorithms offer strong generalization in theory,” he explains, “but in practice, they require carefully designed spaces and significant manual effort to function effectively.” His curiosity led him to explore the intersection of symbolic planning and machine learning—two traditionally distinct approaches in robotics. 

The core idea behind his project is deceptively simple: enable robots to reason more like humans. When we put clothes in a closet,

 “We don’t consciously model every object or constraint,” Chamzas says. “We just follow an abstract plan: go to the closet, open the door, put the clothes inside.” But for a robot, that same task requires a detailed, manually encoded model. His research seeks to change that by allowing robots to learn abstract representations of tasks and actions directly from experience, rather than relying on human-specified models.   

Technically, the project focuses on enabling robots to perform long-horizon manipulation tasks by learning symbolic abstractions from real-world data.  “Instead of assuming a perfect model of the world,” Professor Chamzas states, “the robot will autonomously collect and analyze its own experience to discover how to represent tasks and actions symbolically.” The result enables more adaptive and explainable robotic behavior.   

The grant application process, Professor Chamzas notes, was both challenging and rewarding. “The CRII program is unique in how it supports early-career researchers,” he says. It gave him the “opportunity to distill my long-term research vision into a focused, high-impact proposal.” He credits the support of his colleagues in the Robotics Department and past CRII recipients for helping him refine his ideas and navigate the application process.    

For other researchers seeking NSF funding, Professor Chamzas offers practical advice: “Start early and don’t be afraid to share your ideas with trusted peers and mentors. Treat the proposal not just as a funding opportunity, but as a chance to clarify and articulate your long-term research vision” He emphasizes the importance of grounding proposals in prior work and being open to feedback—even when it’s conflicting. Professor Chamzas says that open dialogue with colleagues is what helped him the most, and he strongly encourages open conversations.    

With this NSF grant, Professor Chamzas is poised to make significant strides in the field of robotics, pushing the boundaries of how machines learn, plan, and interact with the world. The work supported by this award will contribute to broader developments in the field and provide valuable insights for the robotics community at large.  

Latest Announcements - Talent & Inclusion