Random thoughts on difficult problems
This proposal is in multiple languages, click here to see it in Italian
Explore how random choices can tackle tough problems! Without using any special library, we’ll solve challenges with the simple yet powerful ‘random walk’ method - starting with the trivial and venturing into the impossible. Code only… no advanced math or Greek letters required!
In this talk, I’ll introduce random walk optimization techniques with a practical approach. By solving concrete optimization problems (starting from trivial ones and moving to more challenging cases) I’ll demonstrate how random walk methods can be easy to implement while providing reasonably good answers.
Of course, there’s no silver bullet, and much depends on the details (as always!). Nevertheless, it’s fascinating how randomness can often turn out to be an effective strategy.
The talk will feature a live coding session using Jupyter notebooks. While the concepts aren’t strictly Python-specific, Python is an excellent tool for quickly testing ideas before committing to a more performance-oriented implementation in, say, C++ to fully leverage [CG]PU(s).
Andrea Griffini
Coding since 1981, still having fun coding. Math degree but always worked in coding in many different fields (from business software to video games, from CAD to artificial vision, from robotics to network protocols, from compilers to machine learning…). Used visual studio for many years, then vi for many years, then emacs for many years. Currently using my own editor. Recently main occupation is on artificial vision and nesting optimization problems.
Scrivo codice dal 1981, e ancora mi diverto. Laurea in matematica, ma ho sempre scritto codice. Ho lavorato in svariati campi (dai gestionali ai video games, da CAD a visione artificiale, da robotica a protocolli di rete, da compilatori a machine learning…) Ho usato visual studio per molti anni, quindi vi per molti anni, quindi emacs per molti anni. Ora sto usando un mio editor. Attualmente le mie occupazioni principali sono visione artificiale e ottimizzazione nesting.