DOWLOAD-DESKTOP-NOTEBOOK.nb (1.04 MB)
– you have earned Featured Contributor Badge ![]()
Your exceptional post has been selected for our editorial column Staff Picks http://wolfr.am/StaffPicks and Your Profile is now distinguished by a Featured Contributor Badge and is displayed on the Featured Contributor Board. Thank you!
Thanks Brian for your brilliant exposition of Bayesian optimization using recurrence and comparison with alternate methods of linear regression. I like your warnings about using matrix multiplication, which is always the slowest part in many algorithms, especially with ML. However DeepMind’s AI program AlphaTensor has recently discovered quicker methods than Strassen’s algorithm, see AI Reveals New Possibilities in Matrix Multiplication | Quanta Magazine
The Deep-Mind results are excellent! Thank you! They used reinforcement learning to find their algorithms. In an amusing twist, my little article has a recurrent approach to Policy-Gradient that could sidestep some matrix multiplication in RL itself.