Libero
PulseAugur coverage of Libero — every cluster mentioning Libero across labs, papers, and developer communities, ranked by signal.
- developed World-Action Models 90%
- developed by Fast-WAM 90%
- developed Fast-WAM 90%
- instance of World-Action Models 90%
- used by RoboTwin 2.0 70%
- developed LIBERO-Plus 70%
- instance of LIBERO-Plus 70%
- used by LIBERO-Plus 70%
- used by World-Action Models 70%
- used by Vision Language Action (VLA) models 70%
- instance of Openvla 70%
- instance of roboTwin 70%
7 day(s) with sentiment data
-
New method steers generative robot policies with prioritized objectives
Researchers have developed a novel method to steer pre-trained generative robot policies at inference time, allowing them to adhere to prioritized deployment objectives without altering the policy's weights. This approa…
-
New research enhances robot manipulation with temporal context and visual foresight · 4 sources tracked
Researchers are developing new methods to improve robot manipulation by incorporating temporal context and visual foresight. PACT-WAM uses compact temporal encoding to predict action trajectories and visual outcomes, ac…
-
Embodied AI model DM0.5 sweeps four key benchmarks, showing broad capability
The embodied AI model DM0.5 has achieved top rankings across four sub-leaderboards on the RoboColiseum benchmark, a comprehensive evaluation platform for embodied intelligence. This model is the first to excel in all fo…
-
New research reframes diffusion model optimization for reinforcement learning
Researchers have proposed new methods for optimizing diffusion models, particularly in the context of reinforcement learning. One approach, detailed in "Freeze, Share, Shrink," suggests that the action backbone in diffu…
-
AtomicVLA framework enhances robotic skill learning with atomic abstraction
Researchers have introduced AtomicVLA, a novel framework designed to enhance the capabilities of Visual-Language-Action (VLA) models in robotics. This system addresses the limitations of current monolithic VLA models by…
-
New CASD method enhances robot manipulation by distilling semantic targets
Researchers have developed a new method called Chunk-Aligned Semantic Distillation (CASD) to improve robot manipulation tasks. CASD uses an offline vision-language model to segment demonstrations into stages and derive …
-
New frameworks advance AI world models for simulation and robotics · 8 sources tracked
Several research papers introduce new frameworks and models for advancing world-action models, which are crucial for enabling AI agents to understand and interact with the physical world. World in World offers a trainin…
-
New World Model Trains Robots for Better Performance by Withdrawing Post-Training · 1 source tracked
Researchers have developed a novel world model, Phi-WM 1.0 ActEffect, designed to improve robot performance by strategically withdrawing from the deployment process after training. Unlike traditional models that remain …
-
Apple's REFACTOR-VLA learns reusable skills for robots
Apple's REFACTOR-VLA system addresses limitations in current vision-language-action (VLA) models by learning reusable skills through a wake/sleep architecture. Unlike monolithic models that output raw commands, REFACTOR…
-
AcrossWAM1.0 modularizes robot policy stack, reducing parameters with minimal performance loss
Researchers have developed AcrossWAM1.0, a modularized version of the LaWAM framework for robot policies. This new approach separates the world model, multimodal backbone, and deployment checkpoint, allowing for more au…
-
StreamPI framework gives VLA models temporal understanding of physical world
Researchers from Da Xiao Robotics and the University of Hong Kong have introduced StreamPI, a novel framework designed to imbue Vision-Language-Action (VLA) models with a temporal understanding of the physical world. Un…
-
AI research focuses on efficient visual token processing for transformers
Researchers are developing new methods to optimize the use of visual tokens in AI models, particularly for vision transformers and vision-language models. These approaches aim to reduce computational costs and improve i…
-
New RARM method boosts robot manipulation RL with single demo
Researchers have developed a new method called RARM (Reference-Anchored Reward Model) to improve reinforcement learning for robot manipulation tasks. RARM uses a single successful demonstration to create a progress-awar…
-
Yuanli Lingji's DM0.5 Embodied AI Model Achieves SOTA, Open-Sourced
Yuanli Lingji's DM0.5 model has achieved state-of-the-art performance on the RoboDojo benchmark for embodied AI, demonstrating exceptional memory capabilities and a 19.34% average success rate. The model also excels in …
-
New World Action Models Enhance AI Agents in Games and Robotics
Researchers are developing novel World Action Models (WAMs) to improve agent performance in video games and robotics. GameWAM, for instance, unifies visual prediction and action generation for native game control, demon…
-
New QWM Framework Enhances Reinforcement Learning with World Models
Researchers have introduced QWM, a novel framework that integrates world models with Q-learning to enhance sample efficiency in reinforcement learning. This approach uses world models for test-time search over imagined …
-
SpecVLA framework enhances VLA model efficiency for embodied AI
Researchers have developed SpecVLA, a novel framework for co-designing algorithms and hardware architectures to improve the efficiency of Vision-Language-Action (VLA) models in embodied AI. This approach leverages the o…
-
GaussMemory: New 3D Gaussian Splatting for Task-Driven Robotic Memory
Researchers have introduced GaussMemory, a novel approach to robotic manipulation that utilizes 3D Gaussian Splatting for task-driven spatial memory. Unlike previous passive systems, GaussMemory actively learns which ob…
-
G0.5 model integrates robot reasoning and action in single stream
Researchers have introduced G0.5, a novel autoregressive Vision-Language-Action (VLA) model that integrates reasoning and action generation within a single Transformer decoder. This approach allows the VLM to act as a d…
-
RIFT method slashes robotic action latency by removing iterative video rollout
Researchers have developed RIFT (Rollout-free Imagination via Future Tokens), a novel method for World Action Models (WAMs) that significantly reduces latency by eliminating iterative video rollout. By using learned ant…