[ad_1]
The laws of physics have been developed through observation and collection of data that characterize a phenomenon or an event. Typically mathematical equations are developed and fit to match the observed data. The math generally is able to describe simple systems fairly accurately.
Similarly, the path that machine learning takes is also based on looking at and identifying patterns in data. Researchers are trying to apply machine to help learn and improve out knowledge of the laws of physics.
For example, researchers at Purdue have developed a tool that uses machine learning that has been used to “rediscover” Newton’s second law of motion and Lindemann’s Law, a method for predicting the melting temperature of materials. Researchers at DeepMind have similarly used machine learning as a tool to model and predict chemical and physical properties of materials.
Machine learning and physical laws modeled with mathematics can be complementary tools.
Ömer Özgür, in an article for TowardsAI, wrote that “if you train a deep learning model to predict the falling of apples probably, it won’t be helpful to use this model in space missions. Deep learning is not the solution to every problem. Deep learning proves extraordinarily efficient at learning in high-dimensional spaces but suffers from poor generalization and interpretability. On the other hand, Symbolic Regression is very good at generalization, but it isn’t very good at high-dimensional data.”
[ad_2]
Source link