CLAW @ CMU

By popular request, I’ve decided to occasionally dump some links here.

Planning

• LLM Powered Autonomous Agents (blog)
• LaMPP: Language Models as Probabilistic Priors for Perception and Action
• LLM+P: Empowering Large Language Models with Optimal Planning Proficiency
• Do Embodied Agents Dream of Pixelated Sheep?: Embodied Decision Making using Language Guided World Modelling
• Grounding Classical Task Planners via Vision-Language Models

Navigation

• ViNT: A Foundation Model for Visual Navigation
• SACSoN: Scalable Autonomous Data Collection for Social Navigation
• A System for Generalized 3D Multi-Object Search
• CoWs on Pasture: Baselines and Benchmarks for Language-Driven Zero-Shot Object Navigation
• Principles and Guidelines for Evaluating Social Robot Navigation Algorithms

Manipulation

• RVT: Robotic View Transformer for 3D Object Manipulation
• Physically Grounded Vision-Language Models for Robotic Manipulation
• LATTE: LAnguage Trajectory TransformEr
• LIV: Language-Image Representations and Rewards for Robotic Control
• Gesture-Informed Robot Assistance via Foundation Models
• SPRINT: Semantic Policy Pre-training via Language Instruction Relabeling
• Language-Driven Representation Learning for Robotics
• EmbodiedGPT: Vision-Language Pre-Training via Embodied Chain of Thought
• VoxPoser: Composable 3D Value Maps for Robotic Manipulation with Language Models
• VIMA: General Robot Manipulation with Multimodal Prompts
• Using Both Demonstrations and Language Instructions to Efficiently Learn Robotic Tasks

Mobile Manipulation

• Grounded Decoding: Guiding Text Generation with Grounded Models for Robot Control
• Open-World Object Manipulation using Pre-Trained Vision-Language Models
• Robotic Skill Acquisition via Instruction Augmentation with Vision-Language Models
• HomeRobot: Open-Vocabulary Mobile Manipulation
• LSC: Language-guided Skill Coordination for Open-Vocabulary Mobile Pick-and-Place
• Spatial-Language Attention Policies
• TidyBot: Personalized Robot Assistance with Large Language Models
• SayPlan: Grounding Large Language Models using 3D Scene Graphs for Scalable Task Planning

Language to Motion

• Language to Reward for Robot Skill Synthesis
• Text2Motion: From Natural Language Instructions to Feasible Plans
• SayTap: Language to Quadrupedal Locomotion

Multi-Platform

• ChatGPT for Robotics: Design Principles and Model Abilities

Dialogue and QA

• SEAGULL: An Embodied Agent for Instruction Following through Situated Dialog
• SQA3D: Situated Question Answering in 3D Scenes
• Robots That Ask For Help: Uncertainty Alignment for Large Language Model Planners

Other

• Large Language Models as General Pattern Machines
• Language to Rewards for Robotic Skill Synthesis
• Where are we in the search for an Artificial Visual Cortex for Embodied Intelligence?
• MotionGPT
• Modeling Dynamic Environments with Scene Graph Memory
• Affordances from Human Videos as a Versatile Representation for Robotics
• RoboCat: A self-improving robotic agent
• ImageNet-Hard: The Hardest Images Remaining from a Study of the Power of Zoom and Spatial Biases in Image Classification
• Language Is Not All You Need: Aligning Perception with Language Models
• Kosmos-2: Grounding Multimodal Large Language Models to the World
• Behavior Transformers: Cloning k modes with one stone
• From Play to Policy: Conditional Behavior Generation from Uncurated Robot Data
• Affordance Diffusion: Synthesizing Hand-Object Interactions
• Robot Learning on the Job: Human-in-the-Loop Autonomy and Learning During Deployment
• https://sites.google.com/view/robot-r3m/