Luke Zettlemoyer

About

Luke Zettlemoyer is a prominent researcher at the intersection of natural language processing and robotics, with a particular focus on language grounding, human-robot interaction, and embodied AI. His work addresses one of the field's most compelling challenges: enabling untrained users to communicate naturally and intuitively with robotic systems operating in real-world environments. Among his most influential contributions is his research on grounded attribute learning, which models the relationship between natural language and visual perception, amassing over 270 citations across related publications. His 2013 work on parsing natural language commands for robot control systems (327 citations) demonstrated how robots could interpret complex linguistic instructions and act upon them reliably. Zettlemoyer has also pioneered research into deictic gesture and language learning, exploring how humans naturally combine gesture and speech during unscripted robot interactions, work that has accumulated over 150 citations combined. More recently, his Vision-and-Dialog Navigation dataset (119 citations) pushed boundaries by enabling robots to engage in cooperative dialogue while navigating photorealistic environments. His 2021 work extending language grounding to 3D objects reflects his continued commitment to bridging the gap between language understanding and physical-world perception, making robots genuinely useful collaborators in everyday human spaces.

Research Focus

Key Achievements

Top Papers

1. 1
   Learning to Parse Natural Language Commands to a Robot Control System

   327 citations · 2013
2. 2
   A Joint Model of Language and Perception for Grounded Attribute Learning

   184 citations · 2012
3. 3
   Vision-and-Dialog Navigation

   119 citations · 2019
4. 4
   A Joint Model of Language and Perception for Grounded Attribute Learning

   90 citations · 2012
5. 5
   Learning from Unscripted Deictic Gesture and Language for Human-Robot Interactions

   83 citations · 2014
6. 6
   Learning from Unscripted Deictic Gesture and Language for Human-Robot Interactions

   70 citations · 2014
7. 7
   Robot Programming by Demonstration with situated spatial language understanding

   65 citations · 2015
8. 8
   Language Grounding with 3D Objects

   17 citations · 2021
9. 9
   Learning and Planning with Probabilistic Relational Rules

   5 citations · 2004
10. 10
    Combining World and Interaction Models for Human-Robot Collaborations

    3 citations · 2013

Key Collaborators

Contact & Links