Deepak Pathak
Associate Professor
Studies robot learning and self-supervised learning, including curiosity-driven exploration and generalist robot policies.
Deepak Pathak's lab at CMU works on manipulation & robot learning and generative & foundation models. 38 papers since 2023, 35 at top venues, mostly ICML and CoRL. 9 PhD students and 4 master's students. 24 alumni, most often to Skild AI and Tesla. Applications close .
At a glance
- papers, 2023–2026
- 10/ year
- at T1 venues
- 92%
- PhD students
- 9
- MS students
- 4
- alumni listed
- 24
Papers by yearfrom the lab homepage, 2023–2026
- 5 papers2026 · T1 5 · collecting
- 7 papers2025 · T1 7
- 10 papers2024 · T1 9
- 16 papers2023 · T1 14
T1 venuesOtherVenues: ICML, CoRL, ICRA, CVPR, ICLR
Research focuspaper titles that use each topic's words
- Manipulation & robot learning13 papers
- Generative & foundation models10 papers
- robot4 papers
- generalist2 papers
- robot policies1 paper
Paper list
From the lab homepage.
20265 papers
- ICML (T1)3D-DLP: Self-supervised 3D Object-centric Scene Representation Learning (opens in a new tab)
- ICML (T1)Expanding the Capabilities of Reinforcement Learning via Text Feedback (opens in a new tab)
- ICLR (T1)Latent Particle World Models: Self-supervised Object-centric Stochastic Dynamics Modeling (opens in a new tab)
- ICML (T1)Solving Physics Olympiad via Reinforcement Learning on Physics Simulators (opens in a new tab)
- ICLR (T1)ViPRA: Video Prediction for Robot Actions (opens in a new tab)
20257 papers
- RSS (T1)DexWild: Dexterous Human Interactions for In-the-Wild Robot Policies (opens in a new tab)
- NeurIPS (T1)Diffusion Beats Autoregressive in Data-Constrained Settings (opens in a new tab)
- RSS (T1)FACTR: Force-Attending Curriculum Training for Contact-Rich Policy Learning (opens in a new tab)
- ICLR (T1)Generative Classifiers Avoid Shortcut Solutions (opens in a new tab)
- ICRA (T1)Local Policies Enable Zero-shot Long-horizon Manipulation (opens in a new tab)
- CoRL (T1)LocoFormer: Generalist Locomotion via Long-context Adaptation (opens in a new tab)
- IROS (T1)Neural MP: A Generalist Neural Motion Planner (opens in a new tab)
202410 papers
- CoRL (T1)Bimanual Dexterity for Complex Tasks (opens in a new tab)
- ECCV (T1)Evaluating Text-to-Visual Generation with Image-to-Text Generation (opens in a new tab)
- ICRA (T1)Extreme Parkour with Legged Robots (opens in a new tab)
- CVPR @ SynData WorkshopGenAI-Bench: Evaluating and Improving Compositional Text-to-Visual Generation (opens in a new tab)
- CVPR (T1)Language Models as Black-Box Optimizers for Vision-Language Models (opens in a new tab)
- ICLR (T1)Meta-Evolve: Continuous Robot Evolution for One-to-many Policy Transfer (opens in a new tab)
- ICRA (T1)Open X-Embodiment: Robotic Learning Datasets and RT-X Models (opens in a new tab)
- ICML (T1)Revisiting the Role of Language Priors in Vision-Language Models (opens in a new tab)
- ICML (T1)SAPG: Split and Aggregate Policy Gradients (opens in a new tab)
- CVPR (T1)SPIN: Simultaneous Perception, Interaction and Navigation (opens in a new tab)
202316 papers
- CVPR (T1)Affordances from Human Videos as a Versatile Representation for Robotics (opens in a new tab)
- ICRA (T1)ALAN : Autonomously Exploring Robotic Agents in the Real World (opens in a new tab)
- IEEE-RAS HumanoidsDASH: A Framework for Designing Anthropomorphic Soft Hands through Interaction (opens in a new tab)
- CoRL (T1)DEFT: Dexterous Fine-Tuning for Real-World Hand Policies (opens in a new tab)
- CoRL (T1)Dexterous Functional Grasping (opens in a new tab)
- NeurIPS (T1)Diffusion-TTA: Test-time Adaptation of Discriminative Models via Generative Feedback (opens in a new tab)
- ICML (T1)Efficient RL via Disentangled Environment and Agent Representations (opens in a new tab)
- WACVFLAVR: Flow-Agnostic Video Representationsfor Fast Frame Interpolation (opens in a new tab)
- ICML (T1)Internet Explorer: Targeted Representation Learning on the Open Web (opens in a new tab)
- RSS (T1)LEAP Hand: Low-Cost, Efficient, and Anthropomorphic Hand for Robot Learning (opens in a new tab)
- ICRA (T1)Legs as Manipulator: Pushing Quadrupedal Agility Beyond Locomotion (opens in a new tab)
- CVPR (T1)Multimodality Helps Unimodality: Cross-Modal Few-Shot Learning with Multimodal Models (opens in a new tab)
Where alumni went24 listed, 39% to industry
Industry 9Academia 14
Most often to
- Skild AI
- Tesla
- Meta
Also to
- MIT
- Princeton University
- Stanford University
Apply to CMUFall 2027
CMU SCS Computer Science PhD
- Round
- Single round
- Opens
- GRE
- See program page
- English
- TOEFL / IELTS / Duolingo (F-1/J-1 non-native speakers)
- Fee
- USD 100 (USD 80 by early deadline)
Confirmed on the program page.
Similar labsnearest by research summary
- USStanford
Chelsea Finn
- Manipulation & robot learning
- SGNTU
Wei Xiao
- Planning, control & multi-robot systems
- USUW
Byron Boots
- Manipulation & robot learning
- SGNUS
Yuejiang Liu
- Generative & foundation models
- USGeorgia Tech
Glen Chou
- Planning, control & multi-robot systems
- USMIT
Leslie Kaelbling
- Planning, control & multi-robot systems
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.
CMUcloses