Google apuesta por la simulación del mundo real con un nuevo equipo de IA en DeepMind

DeepMind has announced the creation of a specialized team focused on developing “world models,” artificial intelligence systems designed to simulate physical and virtual environments with a level of precision indistinguishable from reality. These large-scale generative models aim ultimately to accelerate progress toward Artificial General Intelligence (AGI), enabling machines to understand and predict the behavior of the physical world through pretraining on visual and multimodal data. The significance of this technology lies in its ability to transform key industries such as robotics, video games, and interactive entertainment. By allowing autonomous agents to be trained in safe simulations, it reduces the risks and costs associated with testing in real-world environments. This facilitates the development of smarter robots capable of adapting to complex situations, while also enhancing the level of immersion and realism in digital experiences. This advancement is crucial for the open_data ecosystem because it democratizes access to high-fidelity simulations. By generating large volumes of accurate synthetic data, these models provide essential datasets for training computer vision and physical reasoning systems without relying exclusively on real-world data collection. This accelerates innovation in open AI, ensuring that researchers and developers can test algorithms in realistic and ethical contexts, fostering transparency and collaboration in the development of safe and robust artificial intelligence.

Source: wwwhatsnew.com
Published on 2025-01-09

Experta en seguridad advierte sobre práctica con paquetes de pedidos en línea: evite fraude - Pulzo »

Open Data

Related news