6 Life Lessons I Learned From AI Research

Подписаться 1,6 млн

Просмотров 120 тыс.

50% 1

Tensorflow experiment link: / d20noqu
Karpathy’s classifier neural network: cs.stanford.edu/people/karpat...
Pick up cool perks on our Patreon page:
› / twominutepapers
We would like to thank our generous Patreon supporters who make Two Minute Papers possible:
313V, Alex Haro, Andrew Melnychuk, Angelos Evripiotis, Anthony Vdovitchenko, Brian Gilman, Christian Ahlin, Christoph Jadanowski, Dennis Abts, Eric Haddad, Eric Martel, Evan Breznyik, Geronimo Moralez, Jason Rollins, Javier Bustamante, John De Witt, Kaiesh Vohra, Kjartan Olason, Lorin Atzberger, Marcin Dukaczewski, Marten Rauschenberg, Maurits van Mastrigt, Michael Albrecht, Michael Jensen, Morten Punnerud Engelstad, Nader Shakerin, Owen Campbell-Moore, Owen Skarpness, Raul Araújo da Silva, Richard Reis, Rob Rowe, Robin Graham, Ryan Monsurate, Shawn Azman, Steef, Steve Messina, Sunil Kim, Thomas Krcmar, Torsten Reil, Zach Boldyga, Zach Doty.
/ twominutepapers
Thumbnail background image credit: pixabay.com/photo-1807526/
Splash screen/thumbnail design: Felícia Fehér - felicia.hu
Random walk image source: en.wikipedia.org/wiki/Random_...
Tensorflow experiment link: / d20noqu
Karpathy’s classifier neural network: cs.stanford.edu/people/karpat...
Károly Zsolnai-Fehér's links:
Facebook: / twominutepapers
Twitter: / karoly_zsolnai
Web: cg.tuwien.ac.at/~zsolnai/

Опубликовано:

14 янв 2019

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 296

@kashmira_zambad 5 лет назад

Summary in two-minutes:

@factsheet4930 5 лет назад

You should also remember, that when an AI starts his first baby steps, it's dumber than anyone who has ever lived... but through determination, trail and error, and the inability to give up, it finds its way up and eventually preforms better than any human would ever do learning from its previous mistakes :)

@AndioDAndia 3 года назад

I base my probability of success in this formula: S=1-(1-s)^i, where S is the final probability of success, s is the average probability of success of all intents, and i is the number of intents. Let's say my average probability of success is 0.1% (s=10^-3). If I try, say 3 times (i=3), my final probability of success is 0.3% (S=3·10^-3). But if I try a thousand times (i=1000), it climbs to 63% (S=0.63). So all I have to do is try over and over again, and my success is garanteed. So yes, intelligence leads to success, inevitably.

@TwoMinutePapers 5 лет назад

The longest ever Two Minute Papers episode that's not two minutes and it's not about a paper (but many). You know, just the usual. Hope you Fellow Scholars enjoyed it! :)

@briancrane7634 5 лет назад

(1) have a

@heartbeatpoetry7704 5 лет назад

I think there needs to be a corollary added to the "if you find something good hang onto it" lesson. You don't want to get yourself stuck in a local maxima. What I learned from it is sometimes injecting a bit of randomness or chaos into your life can shake things up and get you out of the local maxima. The key is finding the sweet spot, too little and it doesn't help, too much and you are starting from scratch again.

@deadpoolbr0 5 лет назад

That day has come, AI teaches people how to live.

@bieeebersarah9078 4 года назад

my problem is that my pleasure comes form doing thing with high efficience, to be in a flux state. my own loss function of my brain tells me i have to change my path when i move little forward in something i have put large amount of time... these advice sounds a bit like a "follow your dreams and hold on no matter what" but with more words. other than that still loves your channel keep it up

@iVideoCommenter 5 лет назад

I think you'll enjoy a book also about life lessons and AI written by experts on evolutionary algorithms called Why Greatness Cannot Be Planned: The Myth of the Objective (Ken Stanley, Joel Lehman). Its main point is that your first point only works for minor goals, but not ambitious goals! Objectives/measuring progress is actually detrimental to achieving ambitious goal because the stepping stones do not resemble the finished product. It uses many analogies with both evolutionary algorithms and actual evolution in nature (i.e. natural selection) which both are capable of greatness and yet usually have a non-obvious utility function,

@MarcelToma 5 лет назад

This video touched me and pulled me from the lack of progress I had. It just showed, objectively, that I'm progressing in my studies, I just have to keep focus and persist

@DamianReloaded 5 лет назад

Fell in a local minima? Get wasted and try again! (Or change your hidden layers!)

@MobyMotion 5 лет назад

Great video! I love what AI can teach us about learning in ourselves. I.e. the same techniques we can use to make machines learn effectively, can be used when we revise for exams. Revising diagnoses by thinking of concrete “examples”, like specific patients I’ve met with that condition, makes learning them so much easier.

@kvackkvack 5 лет назад

Your genuine joy of talking about the paper in question hits me with so much motivation and curiosity! I just want to explore all of the possibilities that popup in my head. What a time to be alive, thanks!

@thesuccessfulone 5 лет назад

Hey man, I really love that you did this. I appreciate you aggregating a bunch of life advice from AI research. It's an interesting parallel between how our minds work and how we achieve our goals and how AI and mathematical experiments achieve the goals of the papers.

@c64cosmin 5 лет назад

This new format is actually awesome, great input from people that I follow. Would be great would be to gather this kind of advices from your fellow colleagues or researchers around the world. I guess the channel could grow even bigger. Kudos!

@abisz007007 5 лет назад

that where the longest 2 min i have ever experienced. And i enjoyed every second of it. Lllllove it

@Eschelaun 5 лет назад

I loved the message behind this, thank you for your inspiring videos.

@taukapow Год назад

Thank you Károly, your videos are exciting to watch and stir the curiosity. Keep up the great work.

@sarthakmishra1415 5 лет назад

This is simply beautiful

@Soulthym 5 лет назад

I totally relate to all those people who said they were inspirer by your work. And dang am I too. Without you I wouldn't have discovered half of what I know about AI, so much inspiring!