Joyce Chai
Professor, EECS – Computer Science and Engineering
Studies natural language processing grounded in vision and robotics, including situated human-machine communication and interactive task learning.
Joyce Chai's lab at Michigan works on multimodal & grounded language and signal processing & information theory. 49 papers since 2023, 34 at top venues, mostly ACL and EMNLP. 5 PhD students. 88 alumni. Applications close .
At a glance
- papers, 2023–2026
- 13/ year
- at T1 venues
- 69%
- PhD students
- 5
- MS students
- 0
- alumni listed
- 88
Papers by yearfrom the lab homepage, 2023–2026
- 6 papers2026 · T1 5 · collecting
- 17 papers2025 · T1 10
- 13 papers2024 · T1 10
- 13 papers2023 · T1 9
T1 venuesOtherVenues: ACL, EMNLP, CVPR, Findings of EMNLP, ICRA
Research focuspaper titles that use each topic's words
- Multimodal & grounded language24 papers
- grounded4 papers
- Signal processing & information theory3 papers
- task0 papers
- human-machine0 papers
Paper list
From the lab homepage.
20266 papers
- CoLMCommunication and Verification in LLM Agents towards Collaboration under Information Asymmetry
- ECCV (T1)Fast Spatial Memory with Elastic Test-Time Training
- ICML (T1)RoboMME: Benchmarking and Understanding Memory for Robotic Generalist Policies
- ACL (T1)SafetyALFRED: Evaluating Safety-Conscious Planning of Multimodal Large Language Models
- ACL (T1)Sparse Feature Coactivation Reveals Causal Semantic Modules in Large Language Models
- ICML (T1)The Mechanistic Emergence of Symbol Grounding in Language Models
202517 papers
- CVPR (T1)3D-GRAND: A Million-Scale Dataset for 3D-LLMs with Better Grounding and Less Hallucination
- NeurIPS (T1)4D-LRM: Large Space-Time Reconstruction Model From and To Any View at Any Time
- CoRL (T1)AimBot: A Simple Auxiliary Visual Cue to Enhance Spatial Awareness of Visuomotor Policies
- NAACL (T1)Babysit A Language Model From Scratch: Interactive Language Learning by Trials and Demonstrations
- Findings of EMNLPBenchmarking and Improving LLM Robustness for Personalized Generation
- COLMBootstrapping Visual Assistant Modeling with Situated Interaction Simulation
- Findings of ACLDo Vision-Language Models Have Internal World Models? Towards an Atomic Evaluation
- ICLR (T1)Do Vision-Language Models Represent Space and How? Evaluating Spatial Frame of Reference under Ambiguities
- CVPR (T1)Fast3R: Towards 3D Reconstruction of 1000+ Images in One Forward Pass
- COGSCIFlexible Physical Problem Solving with Strategy Acquisition and Composition
- EMNLP (T1)Proactive Assistant Dialogue Generation from Streaming Egocentric Videos
- ICRA (T1)RACER: Rich Language-Guided Failure Recovery Policies for Imitation Learning
202413 papers
- IROS (T1)DriVLMe: Enhancing LLM-based Autonomous Driving Agents with Embodied and Social Experiences
- EMNLP (T1)Eliciting In-Context Learning in Vision-Language Models for Videos Through Curated Data Distributional Properties
- WACVGIPCOL: Graph-Injected Soft Prompting for Compositional Zero-Shot Learning
- CVPR (T1)GROUNDHOG: Grounding Large Language Models to Holistic Segmentation
- CVPR (T1)Inversion-Free Image Editing with Language-Guided Diffusion Models
- Adaptive Foundation Models: Evolving AI for Personalized and Efficient LearningLinkGPT: Teaching Large Language Models To Predict Missing Links
- ICRA (T1)LLM-Grounder: Open-Vocabulary 3D Visual Grounding with Large Language Model as an Agent
- NeurIPS (T1)Multi-Object Hallucination in Vision-Language Models
- Automation in ConstructionNatural Language Instructions for Intuitive Human Interaction with Robotic Assistants in Field Construction Work
- EMNLP (T1)Teaching Embodied Reinforcement Learning Agents: Informativeness and Diversity of Language Use
- ICRA (T1)Think, Act, and Ask: Open-World Interactive Personalized Robot Navigation
- ICLR (T1)User Modeling Challenges in Interactive AI Assistant Systems
202313 papers
- Findings of EMNLPCan Foundation Models Watch, Talk and Guide You Step by Step to Make a Cake?
- NeurIPS (T1)CycleNet: Rethinking Cycle Consistent in Text‑Guided Diffusion for Image Manipulation
- EMNLP (T1)From Heuristic to Analytic: Cognitively Motivated Strategies for Coherent Physical Commonsense Reasoning
- EMNLP (T1)Grounding Visual Illusions in Language: Do Vision-Language Models Perceive Illusions Like Humans?
- ACL (T1)Human Inspired Progressive Alignment and Comparative Learning for Grounded Word Acquisition
- ACL (T1)In-Context Analogical Reasoning with Pre-Trained Language Models
- Findings of EMNLPMetaReVision: Meta-Learning with Retrieval for Visually Grounded Compositional Concept Acquisition
- ACL (T1)NLP Reproducibility For All: Understanding Experiences of Beginners
- TMLR (T1)Partition-Based Active Learning for Graph Neural Networks
- Alexa Prize SimBot Challenge ProceedingsSEAGULL: An Embodied Agent for Instruction Following through Situated Dialog
- Findings of EMNLPTowards A Holistic Landscape of Situated Theory of Mind in Large Language Models
- IJCAI (T1)Towards Collaborative Plan Acquisition through Theory of Mind Modeling in Situated Dialogue
Where alumni went88 listed, 0% to industry
Industry 0Academia 16
Also to
- UIUC
- Carnegie Mellon University
- Georgia Tech
Apply to MichiganFall 2027
Michigan CSE Computer Science PhD
- Round
- Single round
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- USStanford
Sanmi Koyejo
- Trustworthy & safe ML
- CNZhejiang
Yueting Zhuang
- Knowledge, QA & retrieval
- USStanford
Tengyu Ma
- Learning theory & optimization
- USGeorgia Tech
Jesse Thomason
- Multimodal & grounded language
- CNSJTU
Gaolei Li
- Trustworthy & safe ML
- USGeorgia Tech
Vijay Ganesh
- Trustworthy & safe ML
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
Michigancloses