Tech Meridian ← ENTITY INDEX
PROMY MERIDIAN RU

TOPIC · ENTITY #8317

world models

Related event timeline, sources and context from the news index.

EVENT TIMELINE

4

RESEARCH · 1 SOURCE · arXiv cs.AI

Building World Models from Intrinsic Motivation (arXiv:2611.13113v1)

This paper integrates 24 intrinsic objectives (including curiosity, entropy, and competency variants) into a common model-based RL framework and evaluates the resulting world models via zero-shot, few-shot, and multi-task control across locomotion, manipulation, and open-ended gaming environments using three representation strategies: observation reconstruction, latent prediction, and redundancy reduction. The authors report that curiosity-driven exploration paired with observation reconstruction yields the strongest aggregate performance, while episodic and lifelong novelty signals improve zero-shot transfer and few-shot adaptation in compositional vision-based tasks, motivating joint design of intrinsic motivation and representation learning.

6.0

RESEARCH · 1 SOURCE · arXiv cs.AI

Monograph defines 'Physical AI' framework and surveys techniques (arXiv:2610.27045v1)

A new arXiv monograph (arXiv:2610.27045v1) proposes a unified definition and technical framework for "Physical AI"—systems that perceive, model, and act on the physical world in closed-loop operation—and distinguishes it from related areas (embodied AI, cyber-physical systems, digital twins). The work surveys foundational topics (physics, control, cognition, learning, reasoning), methods (multimodal perception, world models, embodied foundation models, motor control, sim-to-real), applications (robotics, manufacturing, infrastructure, scientific experimentation), and challenges including reliability, real-time execution, safety, evaluation, governance, and long-term deployment.

7.0

MODELS · 1 SOURCE · The Decoder

Reka AI releases Rho-1, a 19B omni-model for text, images, video and robot control

Reka AI published a research preview of Rho-1, a 19-billion-parameter omni-model that processes and generates text, images, continuous video, and robot control actions inside a single neural network using a shared token context without external tool calls. The model was trained on about 320 H100 GPUs over roughly three months and uses an inverse dynamics approach to extract control signals from ordinary internet videos so the same weights can drive both camera prediction and robot movements.

8.0

STARTUPS · 1 SOURCE · The Decoder

Sakana AI hires Jürgen Schmidhuber as Chief Scientific Advisor

Tokyo-based startup Sakana AI has hired Jürgen Schmidhuber as Chief Scientific Advisor to help steer its new RSI lab, which aims to build a self-reinforcing research loop and advance projects like the Darwin Gödel Machine and The AI Scientist. Sakana credits Schmidhuber's early work on deep learning, world models, and meta-learning as influential to its focus on physical, world-model–based autonomous systems.

6.0