[ICML2026] ACTIVE-O3: Empowering Multimodal Large Language Models with Active Perception via GRPO
-
Updated
Apr 30, 2026
[ICML2026] ACTIVE-O3: Empowering Multimodal Large Language Models with Active Perception via GRPO
This repository provides the source code for the paper Reinforcement Learning for Active Perception in Autonomous Navigation.
A Python library for Robotic Information Gathering
Official implementation of Matcha-agent, https://arxiv.org/abs/2303.08268
Deep Reinforcement Learning for Robotic Pushing and Picking in Cluttered Environment
The code release of "Real-time Active Vision for a Humanoid Soccer Robot Using Deep Reinforcement Learning" paper, ICAART 2021
Official code release for our paper 'APPLE: Toward General Active Perception via Reinforcement Learning'.
An active vision system on the PR2 humanoid robot to dynamically detect objects via the head and arm cameras
Learning Underwater Active Perception in Simulation
🤖 Discover the latest research papers on ArXiv tailored to your interests using AI-powered analysis and smart search features.
Minimal Python robotics examples for planning, manipulation, active perception, and embodied AI. No ROS, GPU, or heavy simulator.
The webpage for Matcha-agent project.
GitHub profile for Muhammed Elyamani: structure-aware planning, active sensing, mobile manipulation, SLAM, and robotics control.
Co-GLANCE: Uncertainty-Aware Active Perception for Heterogeneous Robot Teaming
Real-time Uncertainty-Aware Motion Planning for Magnetic-based Navigation (InfoMagNav)
Resource-aware multimodal scene understanding with view selection for efficient captioning and QA.
NeRF-SLAM-oriented neural mapping lab for camera rays, sigma-density learning, opacity/depth rendering, and robotics perception experiments.
A no-signal baseline for visual escalation: reproducible tests of routing signals, visual budgets, and real per-query costs across VLMs.
Energy-based active acquisition of task-aligned driving information channels with a shared generative latent state.
Active evidence assurance for warehouse robots: verify, repair, and gate perception before motion on AMD Radeon and ROCm.
Add a description, image, and links to the active-perception topic page so that developers can more easily learn about it.
To associate your repository with the active-perception topic, visit your repo's landing page and select "manage topics."