Mind's Eye: How physics data improves large language models

Google combines language models with a physics simulator. This is a central fact in the debate about the role of Deep Learning on the way to more general forms of artificial intelligence. Methods like chain-of-thought prompting, more training data and larger models like PaLM led to better results in benchmarks, but not to a fundamental breakthrough. Researchers are therefore experimenting with hybrid approaches that use the language capabilities of AI models to issue queries to specialized, external libraries or systems. Check your inbox or spam folder to confirm your subscription.