Markov Decision Process (MDP) - 5 Minutes with Cyrill

[@krizh289]

boutta fail my AI exam but with your videos atleast I can cram

[@flecko5]

Thanks for the easy explanation!

[@tianyazhang7065]

Like all your videos

[@gunjanshinde396]

At 2:53, it is mentioned that "the policy converges", what exactly we means by 'converges' here ?

[@flecko5]

My understanding is that if you keep updating the policy infinitely, at some point the policy stops changing which means that the policy has "converged".

[@gunjanshinde396]

Hey@@flecko5, thanks! Lately I been realised the same.

[@HamletMeredith]

449 Jermey Loaf

[@WhereToClick]

Thank you!

[@JasonWhite-o7n]

Kuhn Road

[@IvanIvanov-dk6sm]

Thank you for short lecture, but now it is too short. Maybe you could add more equations to make it real 5 min, but not 3.4 min

[@CyrillStachniss]

Thanks for your opinion

[@DDeathdealer007]

@Ivan Ivanov - it is also not 5 minutes, but here is a 50 minute lecture of Cyrill's on this topic: ruclips.net/video/72QwRnSNY88/видео.html

[@TheProblembaer2]

that sounds a little bit entitled. I would rather thank the Professor for taking his time in sharing his knowledge with the world. Much appreaciated.