When I teach AI/ML for executives, I always argue that the intelligence in the name of the filed AI is misleading since what most of the research and practical applications are focusing on is to...
When I teach AI/ML for executives, I always argue that the intelligence in the name of the filed AI is misleading since what most of the research and practical applications are focusing on is to create “rational agents”. Being rational can be achieved with constraint optimization. So do not even expecting anything remotely human intelligence like solutions from AI techniques is a good prior now. However, I also argue that you do not need artificial general intelligence to change fundamentally how our society works. Linear regression - which is the posterchild of curve fitting - changed how we do every almost everything science and engineering related.
I accepted a year or two ago that we'd lost the public battle on calling it machine learning rather than artificial intelligence, but I still think that's a shame. ML is a great descriptive term:...
I accepted a year or two ago that we'd lost the public battle on calling it machine learning rather than artificial intelligence, but I still think that's a shame. ML is a great descriptive term: the machine learns how to do a specific task through repeated trials, rather than being specifically programmed to do it. Still a major step forward, but clearly constrained and suited to a specific class of problems.
When I teach AI/ML for executives, I always argue that the intelligence in the name of the filed AI is misleading since what most of the research and practical applications are focusing on is to create “rational agents”. Being rational can be achieved with constraint optimization. So do not even expecting anything remotely human intelligence like solutions from AI techniques is a good prior now. However, I also argue that you do not need artificial general intelligence to change fundamentally how our society works. Linear regression - which is the posterchild of curve fitting - changed how we do every almost everything science and engineering related.
I accepted a year or two ago that we'd lost the public battle on calling it machine learning rather than artificial intelligence, but I still think that's a shame. ML is a great descriptive term: the machine learns how to do a specific task through repeated trials, rather than being specifically programmed to do it. Still a major step forward, but clearly constrained and suited to a specific class of problems.