Papers

FORCA: A Force-Aware Coding Agent for Contact-Rich Robot Manipulation

Coding agents can translate language instructions into robot programs. For contact-rich manipulation, a key challenge is to generate and refine control logic that responds to physical interaction. We present FORCA, a coding agent that takes visual observations and wrist force/torque (F/T) measurements as input to write Python programs for contact probing and force-responsive motion. Between executions, the agent inspects force records alongside visual outcomes to diagnose failures and revise both motion and force-feedback control logic. FORCA requires no demonstrations, task-specific training, or simulation. Across five real-robot tasks, we demonstrate its effectiveness in three forms of contact-rich manipulation: improving manipulation accuracy, adapting to changes in object or surface properties, and regulating contact under surface variation and object motion. Compared with a vision-only variant of the same agent, FORCA improves grasp accuracy and task success and reduces contact loss in wiping and peeling.

Contact-Rich ManipulationCoding Agent

Xiao Tan, Yixuan Wang, Zhexi Wang · 461 views · Oct 22, 2026

Adaptive Whole-Body MPC with Optimal Posture Planning for Humanoid Robots under Unknown Payloads

When a humanoid lifts an unknown payload, abrupt changes in mass, center of mass, and inertia destabilize whole-body control. The paper couples centroidal MPC, adaptive fuzzy balance compensation driven by capture-point error, and triple-integrator posture-transition planning. MPC produces motion references under contact, friction, and joint constraints; fuzzy compensation allocates corrective torques to stance legs; and the planner generates smooth, feasible sit-to-stand and stand-to-sit trajectories. MuJoCo simulation and a 28-DoF humanoid validate stable lifting with 7 kg and 10 kg payloads and robust 7 kg load-carrying during walking.

HumanoidWhole-Body Control

Manh Cuong Nguyen, Duc Cuong Vu, Viet Anh Vu · 218 views · Oct 9, 2026

DeFM: Learning Foundation Representations from Depth for Robotics

Depth sensors are widely deployed across robotic platforms, and advances in fast, high-fidelity depth simulation have enabled robotic policies trained on depth observations to achieve robust sim-to-real transfer for a wide range of tasks. Despite this, representation learning for depth modality remains underexplored compared to RGB, where large-scale foundation models now define the state of the art. To address this gap, we present DeFM, a self-supervised foundation model trained entirely on depth images for robotic applications. Using a DINO-style self-distillation objective on a curated dataset of 60M depth images, DeFM learns geometric and semantic representations that generalize to diverse environments, tasks, and sensors. To retain metric awareness across multiple scales, we introduce a novel input normalization strategy. We further distill DeFM into compact models suitable for resource-constrained robotic systems. When evaluated on depth-based classification, segmentation, navigation, locomotion, and manipulation benchmarks, DeFM achieves state-of-the-art performance and demonstrates strong generalization from simulation to real-world environments. We release all our pretrained models, which can be adopted off-the-shelf for depth-based robotic learning without task-specific fine-tuning. Webpage: https://de-fm.github.io/

DepthSelf-Supervised Learning

Manthan Patel, Jonas Frey, Mayank Mittal · 166 views · Jan 26, 2026

Learning Multi-Phase Robot Loco-Manipulation via Phase-Conditioned Policy Optimization

Complex whole-body loco-manipulation requires executing diverse maneuvers, such as crouching, grasping, and locomotion, within a single long-horizon task. While performing simple pick-and-place behaviors is well-established, scaling to complex, long-horizon tasks remains difficult. Currently, this challenge is addressed by relying on human motion data and correspondingly humanoid platforms, which limits platform independence and introduces substantial data processing overhead. Instead, we extend proximal policy optimization (PPO) to train a single policy for multi-phase tasks across diverse robotic hardware, leveraging prior task structure to shape behavior. Our method adapts PPO using a phase-conditioned critic, phase-gated reward functions, and a multi-phase reset curriculum to guide exploration. To support varied hardware, we employ a decoupled hierarchical control architecture that separates lower-level balancing dynamics from high-level phase execution. We demonstrate multi-stage loco-manipulation skills on hardware on both the Unitree G1 and a Boston Dynamics Spot equipped with a manipulator arm, and validate our design choices through ablations. Our approach enables the acquisition of complex loco-manipulation behaviors purely through environment interaction while directly incorporating high-level task structure into policy learning.

Whole-Body Loco-ManipulationPhase-Conditioned PPO

Simone Manni, Annika Wollschläger, Dario Bellicoso · 291 views · Oct 26, 2026

EgoScale: Scaling Dexterous Manipulation with Diverse Egocentric Human Data

EgoScale is a human-to-dexterous-manipulation transfer framework. A VLA policy is pretrained on 20,854 hours of action-labeled egocentric human video, revealing a log-linear relationship between data scale and validation loss. Lightweight aligned human-robot mid-training then grounds the learned motor prior in robot control, improving average success by 54% over a no-pretraining baseline on a 22-DoF dexterous hand and transferring to lower-DoF hands.

Dexterous ManipulationVLA

Ruijie Zheng, Dantong Niu, Yuqi Xie · 198 views · Feb 18, 2026

Kernel Forge: An Agent Harness for LLM-based Generation and Optimization of CUDA Kernels

Kernel Forge is an open-source agent harness that accepts an unmodified PyTorch model, captures the operators that execute on representative inputs, and uses an LLM to generate and revise workload-specific CUDA kernels. A persistent Monte Carlo Tree Search controller explores alternative revision paths, while every candidate must compile with nvcc, run on the captured inputs, and match PyTorch eager outputs before it can enter the tree. Exported .cast packages dispatch generated CUDA only when the candidate is valid, compatible, and faster than the captured eager path; otherwise execution falls back to PyTorch. On an NVIDIA DGX Spark GB10, with 50 optimization iterations per kernel, 14 kernels outperform eager across ResNet-50, Stable Diffusion 3.5 Medium, Gemma 4 E2B, and Qwen 3.5 35B-A3B, including 1.52x on adaptive_avgpool2d, 1.70x on group_norm, and 2.83x and 1.54x on softmax.

CUDAGPU Kernel

Joshua Brodsky, Dhravid Kumar, Savini Kashmira · 205 views · Jun 2, 2026

Generative Diffusion Models for Secure UAV-ISAC with Real-Time Target Tracking

A UAV-ISAC system uses a conditional diffusion model to generate message-agnostic safeguard waveforms, hiding event-triggered traffic while jointly optimizing covertness, reliability, and PCRB-based target tracking without a shared key.

UAV-ISACIntegrated Sensing and Communication

Nikbakht, Homa, Eldar, Yonina C., Poor, H. Vincent · 177 views · Oct 20, 2026

Why is AI-generated language so hard to understand?

The authors compare psycholinguistic processing-cost features across 14 language models in chat, coding, and essay settings. Relative to round-trip paraphrases, frontier models overuse low-frequency n-grams, long dependencies, stacked noun compounds, denominal verbs, reduced object relatives, and atypical animacy patterns. They hypothesize that post-training for math, code, and verifiable rewards rewards information density rather than human communication. A Qwen2.5-1.5B Instruct experiment with GRPO shows that a simple length penalty can create a near-categorical preference for shorter variants and transfer to some untrained constructions. Task success therefore does not guarantee human-comprehensible language.

Large Language ModelsPsycholinguistics

Thomas Hikaru Clark, Matias B. Martin, Kyle Mahowald · 330 views · Oct 21, 2026

Self-Adaptive VLA for Robust Robot Deployment

This post-training method turns a policy's own shifted rollouts into latent context, uses a lightweight encoder and adaptive layer normalization to adapt a vision-language-action policy, and recovers over 80% of base performance across four precision-critical manipulation tasks.

VLATest-Time Adaptation

Hongxin Zhang, Chunru Lin, Tsun-Hsuan Wang · 191 views · Sep 24, 2026

FLORA: An Open Model Suite for Scaling Robotics Data Engines

The life-cycle of robotics foundation models crucially depends on data labels. Outcome reward labels are needed for monitoring real-world deployments. Process reward labels are needed during RL post-training. Timestamped captions are needed for filtering and balancing pretraining data. We present Flora, an industrial-quality data engine used for training GR00T models, that simultaneously covers these capabilities. Flora achieves SOTA scores on the captioning benchmark WGO-Bench and reward benchmarks RoboReward and Robometer. On internal benchmarks it is competitive with GPT-6 Astra while beating other frontier models. The model is substantially cheaper than API-based alternatives, roughly 125 times cheaper than GPT-6 Astra. Flora is battle tested at scale, labeling 200,000 hours of data per day across 512 GB300s. By open-sourcing our data engine, we enable researchers to scale custom robotics datasets and explore emerging applications such as physical auto-research and human-to-robot transfer.

Robotics Data EngineTemporal Action Segmentation

Sicong Leng, Yunze Man, Zhiqi Li · 188 views · Oct 22, 2026

InteractiveWorld: A Verifiable Human-Centric Environment for Building Better Interactive MLLM Assistants

People have long dreamed of an assistant that is always there to guide them and remind what they forget, from new dishes to lost keys. Powered by multimodal large language models (MLLMs), egocentric AI assistants on wearable devices are bringing this dream within reach: they see what the user sees and give instructions for user's physical work when necessary. However, testing such assistance is challenging and usually requires real people. The expensive and rare real interaction data leaves a key question unanswered: Can today's MLLMs truly help people in physical world? To answer this question, we propose InteractiveWorld, the first verifiable environment for closed-loop interaction between human and AI assistant. Unlike existing robot-centric simulators, InteractiveWorld lets the assistant guide a human agent who observes the scene, understands instructions, and acts independently. It thus realistically simulates how AI assistant collaborates with human: the assistant instructs, the human acts, and the environment executes actions and updates what they see next. Building on it, we introduce InteractiveBench, the first dynamic benchmark for egocentric AI assistants, with 120 multi-turn tasks across six categories. In each task, the assistant guides the human in real time, and every instruction shapes the outcome, which is scored by verifiable, state-based rewards. Experiments with 12 frontier MLLMs show they are still far from reliable assistants. Even the strongest model, Claude Fable 5.1, achieves at most 20% and 25% on long-horizon Guidance and Planning tasks, respectively. Overall, our work offers a verifiable platform and a rigorous evaluation set, laying a solid foundation for research on interactive MLLM assistants that truly help people in the physical world.

MLLMEgocentric AI Assistant

Qinsi Wang, Wei Wen, Chong Zhou · 225 views · Oct 25, 2026

GloSplat: Joint Pose-Appearance Optimization for Faster and More Accurate 3D Reconstruction

GloSplat unifies global structure from motion initialization with 3D Gaussian Splatting training through joint pose-appearance optimization. Its key design keeps explicit SfM feature tracks as independent optimizable geometry anchors instead of merging them with Gaussian means, so bundle adjustment can correct camera poses without being disrupted by densification or noisy photometric gradients. A retrieval-based variant prioritizes speed, while an exhaustive-matching variant targets maximum reconstruction quality; both improve the accuracy-speed frontier on unposed novel-view-synthesis benchmarks relevant to robot simulation reconstruction.

3D reconstruction3D Gaussian Splatting

Tianyu Xiong, Rui Li, Linjie Li · 128 views · Mar 5, 2026

As an Amazon Associate, we earn from qualifying purchases.