HERO demonstrates self-improving manipulation without human demonstrations
Researchers introduced HERO, a hierarchical embodied agent that autonomously gathers experience and consolidates reusable manipulation policies.
Semiconductor Domain: memory
World-model research identifies limits in learning hidden physical parameters
A controlled study found prediction objectives determine whether robot world models retain mass, stiffness, drag, and force information.
Samsung creates a dedicated robotics organization and international research hubs
Samsung formed a robotics strategy unit and outlined research hubs in the United States, China, and Japan.
Noetra begins full-scale development of a Japan-made multimodal foundation model
Noetra, Sony, SoftBank, NEC, Honda, AIST, and partners started full-scale development of a sovereign multimodal model.
Delta Intelligence raises nearly RMB500 million in angel++ financing
The financing targets foundation-model iteration, data-collection equipment and industrial validation.
On-Device Adaptation Broadens Robot Portability
The on-device model was reported to adapt rapidly across robots with differing sensors and embodiments.
Generalist AI Seeks a Higher Valuation
Generalist AI was reported to be discussing a funding round at a three-billion-dollar valuation.
DLAM Adds Temporal Constraints to Latent Robot Actions
DLAM introduces temporal constraints intended to make latent robot actions more coherent across sequential control decisions.
RedFlow Turns Robot Failures Into Action-Level Corrections
The paper proposes converting failed actions into localized corrections for flow-matching vision-language-action policies.
Action-Conditioned World Models Move Into Planning
The work uses action-conditioned predictive models to improve generalizable embodied decision-making.
