symbolic-mathematics-finally-yields-to-neural-networks-20200520

symbolic-mathematics-finally-yields-to-neural-networks-20200520

6/2/2020

link

https://www.quantamagazine.org/symbolic-mathematics-finally-yields-to-neural-networks-20200520/

summary

This article discusses how researchers have made progress in using neural networks to perform symbolic mathematics, a field that has historically been resistant to machine learning techniques. Symbolic mathematics involves manipulating abstract mathematical expressions symbolically rather than computing numerical values. The article explains the challenges faced in applying neural networks to symbolic mathematics and highlights recent breakthroughs that have enabled neural networks to solve problems involving symbolic reasoning. It also explores the potential applications of this research, such as improving automated theorem proving and advancing artificial intelligence in scientific domains. Overall, the article demonstrates the promising advancements in using neural networks to tackle symbolic mathematical problems.

tags

symbolic mathematics ꞏ neural networks ꞏ machine learning ꞏ artificial intelligence ꞏ mathematics ꞏ computer science ꞏ deep learning ꞏ mathematical modeling ꞏ mathematical concepts ꞏ algorithm ꞏ mathematical equations ꞏ mathematical symbols ꞏ mathematical computation ꞏ mathematical reasoning ꞏ mathematical logic ꞏ mathematical abstraction ꞏ mathematical understanding ꞏ mathematical problem solving ꞏ mathematical representation ꞏ neural computation ꞏ computational neuroscience ꞏ mathematical algorithms ꞏ mathematical patterns ꞏ mathematical analysis ꞏ mathematical complexity ꞏ mathematical learning ꞏ mathematical education ꞏ mathematical cognition ꞏ mathematical thinking ꞏ neural architecture ꞏ mathematical structure ꞏ mathematical prediction ꞏ mathematical optimization ꞏ symbolic computation ꞏ mathematical transformation ꞏ mathematical inference