3 ms·
Im curious about one thing and im self-taugh so please forgive lack of understanding of uni world. Youre saying that physicist who do machine learning should be
by max002 5y ago
Im curious about one thing and im self-taugh so please forgive lack of understanding of uni world. Youre saying that physicist who do machine learning should be instead trying to unserstand physical world. While not having academic background i do read a lot of science news (its my porn:D) and now and then i see stuff like "new strong material discoveded by ml", "ml speeds up discovery of composite materials" or "ml speeds up new drugs discovery" and so on. How that does not improve our understanding of physical world?
- cblconfederate 5y agoPhysics itself is about creating models of the world that are simple and understandable, and provide profound insights. ML is the opposite of that, a black box that works but cannot explain itself. A drug discovered by ML is not increasing our understanding of the world.
- max002 5y agoAnd IF we get to explainable models (i.e. https://aithority.com/machine-learning/neural-networks/deep-learning/abacus-ai-publishes-paper-on-explainable-machine-learning-for-neurips-2021/ https://aithority.com/machine-learning/neural-networks/deep-...) would that change its "usefulness" in physics world in your opinion? Not trying to fight or say that youre wrong. Just curious about perspective, im extremely detail oriented so i pick on them ^_^
- cblconfederate 5y agoit doesnt even need to be explainable. If a model comes up with an equation that reconciles quantum physics with gravity then we could try to reason how it came to it, but we don't need to. As long as the model described in the equation is understandable by humans it should be enough. However most ML models are black boxes that solve problems without generating models of the world (that we know of) Physicists and mathematicians are uncomfortable with computer-generated proofs. I think they 'd be less uncomfortable with computer-generated theories because they can understand those
- max002 5y agoPoints taken, thank you for explanation.
- somesortofthing 5y agoI don't think the author meant to imply that ML as a whole doesn't improve our understanding of the physical world. The point seems to be more that the kind of ML work done by a lot of people who have the qualifications necessary to expand the borders of our understanding of the physical world doesn't meaningfully contribute to our knowledge.