New approach enables physically consistent AI models of chaotic systems

MIT AI News
Read onMIT AI News· 6 min
MIT researchers developed a method to impose physics-based constraints on AI models of chaotic systems, preventing unrealistic predictions in fields like weather and fluid dynamics. This approach enhances model reliability for complex, unstable processes by ensuring outputs stay within physically possible bounds.
New approach enables physically consistent AI models of chaotic systemsSee full coverageFull coverage from 1 sources

More in Machine Learning