PulseAugur / Brief
EN
LIVE 01:27:23

Brief

last 24h
[2/2] 222 sources

Multi-source AI news clustered, deduplicated, and scored 0–100 across authority, cluster strength, headline signal, and time decay.

  1. SIGNIFICANT · 雷峰网 (Leiphone) 中文(ZH) ·

    Xinghai Tu releases new generation embodied foundation model G0.5, comprehensively improving zero-shot generalization capabilities, enabling robots to think and act simultaneously

    Open-source AI lab OpenGalaxea has released G0.5, a new embodied foundation model designed to enable robots to think while acting. This model integrates reasoning and action into a single architecture, allowing robots to perform tasks with zero-shot generalization. G0.5 utilizes a unified action decoder and a native action Chain-of-Thought mechanism, enabling it to decompose long-range tasks and adapt its actions based on natural language prompts. The model also incorporates a spatio-temporal attention module for enhanced contextual awareness, demonstrating state-of-the-art performance on various benchmarks. AI

    Xinghai Tu releases new generation embodied foundation model G0.5, comprehensively improving zero-shot generalization capabilities, enabling robots to think and act simultaneously

    IMPACT Enables robots to perform complex tasks with natural language commands, potentially accelerating adoption of embodied AI in various industries.

  2. TOOL · 量子位 (QbitAI) 中文(ZH) ·

    τ0-WM: The Largest Open-Source Embodied World Model for Pre-training is Here

    Researchers have introduced τ0-World Model (τ0-WM), an open-source embodied world model trained on a massive 30,000 hours of data, with a significant portion (17,800 hours) derived from real robot teleoperation. This model goes beyond predicting future states by incorporating Test-Time Computation, allowing robots to evaluate and select optimal actions before execution, even correcting for potential errors. τ0-WM demonstrates improved performance on complex manipulation tasks compared to previous models, challenging the conventional approach of reserving real-world data solely for fine-tuning. AI

    IMPACT Sets a new precedent for large-scale pre-training with real-world robot data, potentially accelerating embodied AI development.