All papers

The paper optimizes batched task and motion planning to generate high-quality autonomous demonstrations for robotic manipulation training.

The paper compresses video history into evolving lightweight model weights instead of competing for limited input context length.

The paper trains feedforward 3D reconstruction models using privileged motion and gravity cues to improve dynamic scene understanding.

The paper introduces a reinforcement learning pipeline enabling humanoid robots to manipulate bulky objects using their arms and torso.

The paper introduces an adaptive connectivity objective to improve deep learning segmentation of thin, continuous anatomical or physical structures.

The paper introduces a skill-decoupled on-policy distillation method to merge vision-language experts without conflicting gradients or reward hacking.

The paper presents an end-to-end trained multimodal agent that combines visual graphical interface actions and executable code execution.

The paper presents a method for robots to generate targeted preference queries to learn underspecified rewards from humans.

The authors present AVC-Reward, an event-centric framework to evaluate temporal, narrative, and attribute consistency in generated audio-video content.

The authors propose ACEdit, a training and distillation scheme that stabilizes image editing models against accumulating errors across multiple editing turns.

The authors introduce DSP, a framework where robots learn from brief human play and autonomously collect contact-rich data for dexterous manipulation.

The authors introduce 3DCodeVerse, a dataset and multi-agent framework enabling coding agents to generate, reason over, and verify 3D executable code.

The authors introduce GeoState, a view-decoupled 3D world modeling approach separating scene dynamics from camera motion.

The authors present CoDA, a framework where a simulation data engine and a robot action policy co-evolve using memory of past failures.

The authors propose T2-Z, a method optimizing learnable environment latents at test time to help robotic world models adapt to unobserved physical factors.

The authors introduce LIBERO-PeRM, a simulation benchmark for studying how embodied robotic agents learn and follow user preferences in manipulation tasks.

The paper models epileptic dynamics using a multi-stable system to detect critical slowing down and design feedback control for seizure prevention.

The paper examines vulnerabilities of deepfake detection models to backdoor attacks and evaluates related defense strategies.

The paper studies cybersecurity threat detection using sequential deep learning models evaluated on the KDD network traffic dataset.

The paper summary contains webpage metadata rather than research text, though it mentions partial discharge localization for drone inspection using UHF detection.

The paper evaluates machine learning models, finding Random Forest predicts depression in menopausal women with high accuracy.

The paper proposes parallel delay chains to enable low-power neuromorphic circuits to detect anomalies in long electrocardiogram signals.

The paper evaluates machine learning models to detect active network attacks early with minimal preprocessing of raw traffic data.

The paper provides a synthetic dataset to detect passive wireless jamming attacks carried out using reconfigurable intelligent surfaces.

The authors introduce CharXiv, a benchmark of realistic scientific charts designed to test descriptive and reasoning capabilities in multimodal language models.

This paper presents reconstruction inversion techniques and hardware-software co-design to improve multi-patient temporal computed tomography.

This paper presents a computational method to correct for eye motion and optical aberrations in high-resolution retinal imaging.

The paper presents a residual world model to correct simulated actuator dynamics using proprioception to reduce the sim-to-real gap for humanoids.

The paper presents a visuo-inertial system estimating terrain context, locomotion mode, and gait speed for assistive wearable robotics.

The paper proposes a receding-horizon planning method to evaluate future camera viewpoint sequences for inspecting heavily occluded robotic targets.

The paper introduces a GPU-accelerated differentiable contact dynamics framework for simulating coupled deformable and rigid body physical interactions.

The paper presents a benchmark for vision-based 3D reconstruction of submerged sea ice using polar expedition and synthetic simulator data.

The paper introduces a framework that decouples task intent from robot dynamics to enable collective manipulation learning across multiple embodiments.

The paper compares adaptation strategies for vision-language-action models across diverse data scenarios, such as online, offline, and failure trajectories.

The paper introduces a tactile manipulation dataset and a foundation model combining visual planning with force-aware refinement for dexterous robots.

The paper introduces a mode-unified control framework using radial basis function interpolation to continuously control robotic prostheses across varying terrains.

The paper evaluates end-to-end vision-language-action policies for autonomous lunar rover docking under challenging surface lighting conditions.

The paper presents an action-reweighting learning framework for vision-language-action models to learn from human recovery demonstrations in robotic disassembly.

The paper presents a robust model predictive control method using nonlinear reduced-order models that guarantees constraint satisfaction for full-order systems.

The paper develops a method to learn task-relevant constraints from visual demonstrations using sparse 3D keypoints for safe motion planning.

The paper presents probabilistic digital twins that model uncertain physical properties as probability distributions to guide human-robot interaction.

The paper introduces a residual learning framework with motion-aware gating to preserve expressive characteristics during humanoid robot whole-body control.

The paper evaluates an interactive interface connecting a physical master robot controller to a physical simulation using past successful trajectories as guidance.

The paper formulates a geometric refinement algorithm to smooth multi-usage pipe centerlines while maintaining topological clearance in additively manufactured manifolds.

The paper presents a virtual reality training system for vascular intervention combining dual-mode robot control with sensorless interaction estimation.

The paper proposes a navigation framework separating persistent spatial-semantic mapping from task-conditioned high-level decision policies for robots.

The paper examines how vision-language-action policies handle paraphrased instructions and evaluates ensembling and sampling techniques across different prompts.

The paper proposes a method using event-camera observations to selectively correct kinematic predictions for robust spacecraft pose tracking.

The paper introduces a meta-learning method that optimizes simulation parameter distributions to maximize downstream real-world robot manipulation performance.

The paper presents a hierarchical reinforcement learning framework to learn functional dexterous multi-finger robot manipulation directly on real hardware.

This paper presents an open-source procedural generator for creating customizable 2D and 3D environments with line-of-sight visibility analysis.

This paper presents a cooperative dual-arm planning framework that coordinates tabletop wiping with temporary object removal.

This paper introduces a visual roadmap framework built directly on 3D Gaussian Splatting renderings for image-goal navigation.

This paper proposes a world model planning framework to monitor execution and recover robot policies during deployment without retraining.

This paper presents a teleoperation platform and modular coordination framework for learning whole-body mobile manipulation policies.

This paper analyzes failures in mobile manipulation and introduces agentic interventions to coordinate planning, base movement, and manipulation.

This paper presents a simulation benchmark to evaluate wrist-force feedback integration strategies for vision-language-action policies.

This paper presents a morphology-conditioned learning framework to train a single locomotion policy across varied legged robot configurations.

This paper presents an adaptive architecture coordinating exploration, detection, and validation modules for object-goal navigation.

This paper develops an end-to-end differentiable reinforcement learning method for three-dimensional obstacle avoidance in quadrupedal robots.