arXiv Machine Learning
Jun 2
Can Vision Language Models Learn Intuitive Physics from Interaction?
arXiv:2602. 06033v2 Announce Type: replace Abstract: Pre-trained vision language models do not have good intuitions about the physical world.
Learning concepts with energy functions
We’ve developed an energy-based model that can quickly learn to identify and generate instances of concepts, such as near, above, between, closest, and furthest, expressed as sets of 2d points. Our model learns these concepts after only five demonstrations.
Related stories
OpenAI Blog
May 16, 2017
Robots that learn
We’ve created a robotics system, trained entirely in simulation and deployed on a physical robot, which can learn a new task after seeing it done once.
arXiv AI
Jul 3
PhysMani: Physics-principled 3D World Model for Dynamic Object Manipulation
arXiv:2607. 01938v1 Announce Type: cross Abstract: Manipulating fast and dynamically moving targets in unstructured 3D environments remains challenging for embodied AI.
arXiv Machine Learning
Aug 12
Learning in ImaginationLand: Omnidirectional Policies through 3D Generative Models (OP-Gen)
arXiv:2509. 06191v2 Announce Type: replace-cross Abstract: Recent 3D generative models, which are capable of generating full object shapes from just a few images, now open up new opportunities in robotics.
arXiv AI
Jun 2
AffordGen: Generating Diverse Demonstrations for Generalizable Object Manipulation with Afford Correspondence
arXiv:2604. 10579v2 Announce Type: replace-cross Abstract: Despite the recent success of modern imitation learning methods in robot manipulation, their performance is often constrained by geometric variations due to limited data diversity.
arXiv AI
Sep 24
Distillation for Efficient Multitask Manipulation Policies via Conditional Flow Matching
The paper proposes a method to train efficient multi‑task manipulation policies by distilling knowledge from single‑task Conditional Flow Matching (CFM) experts. Instead of training separate models for each task, the authors transfer the experts’ learned velocity fields into a shared policy, combining this distillation signal with the original CFM objective. Experiments on RLBench demonstrate that this approach improves multi‑task performance while keeping the model size fixed, avoiding the need for larger capacity or performance drops seen with naive concatenated training.
arXiv AI
Jun 19
Movement Primitives in Robotics: A Comprehensive Survey
arXiv:2601. 02379v2 Announce Type: replace-cross Abstract: Biological systems exhibit a continuous stream of movements, consisting of sequential segments, that allow them to perform complex tasks in a creative and versatile fashion.
arXiv Computation and Language
Aug 25
Decoupled Physical Modeling and Execution for Physics Reasoning
The paper introduces a framework that separates physical modeling from execution in physics reasoning tasks. It uses a two‑stage post‑training approach: supervised fine‑tuning to build structured models and reinforcement learning with rubric‑based feedback to refine them. Experiments on PhysReason, PhyX, and SeePhys show that this explicit modeling improves reasoning performance by about 3% on average for small LLMs.
Hugging Face Trending Papers
Jun 22
Policy-as-Data: Learning Generalizable HOI Diffusion Models from Simulated Physics
Synthesizing realistic Human-Object Interactions (HOI) is critical for creating embodied avatars and functional virtual environments. However, current data-driven approaches primarily rely on motion capture datasets, which are expensive to scale and limited in functional diversity.
arXiv AI
Sep 21
KnowDemo: Knowledge-Guided Robot Demonstration Generation from Human Videos
KnowDemo is a framework that generates diverse robot demonstrations from human videos by leveraging structured manipulation knowledge. It uses a vision‑language model to extract task requirements and permissible execution variations, then resolves these against target‑scene entities to guide candidate generation and screening before motion planning. The resulting demonstrations feature multimodal behavior, alternative contact strategies, and valid subtask orders, and have been shown to improve planning success and enable sim‑to‑real policy transfer across three tasks.
arXiv AI
Jul 20
SLAC: Safe and Efficient Real-Robot Reinforcement Learning via Unsupervised Simulation Pre-Training
arXiv:2506. 04147v5 Announce Type: replace-cross Abstract: Building capable household and industrial robots requires mastering the control of versatile, high-degree-of-freedom (DoF) systems such as mobile manipulators.
arXiv AI
Sep 24
Intelligence Across Embodiments
The paper discusses how robotic embodiment—sensing, kinematics, dynamics, geometry, actuation, and control—varies across robots and over time, and argues that general embodied intelligence must learn across these differences. It critiques current methods that engineer correspondences for short‑term gains, proposing instead that learning should discover representations that enable transfer across a broader range of embodiments as experience accumulates. The authors advocate for embodiment diversity as a scaling axis, broad learned priors as a complementary ingredient, and evaluations that better characterize embodiment gaps and transfer performance, linking practical cross‑embodiment learning to the scientific pursuit of physical intelligence that adapts with its embodiments.