Yonatan Bisk

About

Yonatan Bisk is a leading researcher at the intersection of natural language processing, robotics, and embodied AI, with a particular focus on grounded language acquisition and human-robot communication. His work addresses one of AI's most fundamental challenges: enabling robots to understand and act upon unrestricted natural language instructions in real-world environments. Bisk's most influential contribution, "Natural Language Communication with Robots" (2016, 104 citations), established a foundational framework for bridging the communication gap between humans and robots, sparking significant follow-on research in the field. He has consistently pushed this agenda forward, exploring how agents can develop recursive mental models for navigating dialogue, reason about 3D objects for more grounded language understanding, and leverage transformer architectures for flexible task planning. More recently, Bisk has turned his attention to foundation models as a pathway to general-purpose robotics, contributing both survey-level meta-analyses and benchmark datasets like OpenEQA (2024, 46 citations) that probe whether modern AI systems can truly understand embodied environments. Projects like HomeRobot further demonstrate his commitment to practical, open-vocabulary manipulation in everyday settings. Across his career, Bisk's research has helped define the roadmap for robots that genuinely comprehend the human world.

Research Focus

Key Achievements

Top Papers

1. 1
   Natural Language Communication with Robots

   104 citations · 2016
2. 2
   OpenEQA: Embodied Question Answering in the Era of Foundation Models

   46 citations · 2024
3. 3
   RMM: A Recursive Mental Model for Dialogue Navigation

   32 citations · 2020
4. 4
   Toward General-Purpose Robots via Foundation Models: A Survey and Meta-Analysis

   26 citations · 2023
5. 5
   Towards a Dataset for Human Computer Communication via Grounded Language Acquisition

   24 citations · 2016
6. 6
   Language Grounding with 3D Objects

   17 citations · 2021
7. 7
   Prospection: Interpretable plans from language by predicting the future

   15 citations · 2019
8. 8
   RMM: A Recursive Mental Model for Dialog Navigation

   14 citations · 2020
9. 9
   HomeRobot: Open-Vocabulary Mobile Manipulation

   13 citations · 2023
10. 10
    Transformers are Adaptable Task Planners

    5 citations · 2022

Key Collaborators

Contact & Links