User Tools

Site Tools


agi

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
agi [2018/12/17 15:24]
admin
agi [2019/06/12 16:08] (current)
admin [THREAD4: Recent Achievements]
Line 129: Line 129:
 Fix Slides are here: http://​people.eecs.berkeley.edu/​~cbfinn/​_files/​mbrl_cifar.pdf Fix Slides are here: http://​people.eecs.berkeley.edu/​~cbfinn/​_files/​mbrl_cifar.pdf
 An older version of this tutorial (1.5 yrs older) is here: https://​www.youtube.com/​watch?​v=iC2a7M9voYU An older version of this tutorial (1.5 yrs older) is here: https://​www.youtube.com/​watch?​v=iC2a7M9voYU
- 
  
 Reinforcement learning is a key part to AGI, but it tends to concentrate on toy problems that can be solved. ​  For a discussion on what's missing from classic reinforcement learning that's needed in AGI: Reinforcement learning is a key part to AGI, but it tends to concentrate on toy problems that can be solved. ​  For a discussion on what's missing from classic reinforcement learning that's needed in AGI:
   * [[https://​www.alexirpan.com/​2018/​02/​14/​rl-hard.html|Deep Reinforcement Learning Doesn'​t Work Yet]]   * [[https://​www.alexirpan.com/​2018/​02/​14/​rl-hard.html|Deep Reinforcement Learning Doesn'​t Work Yet]]
   * [[https://​thegradient.pub/​why-rl-is-flawed/​|Reinforcement learning’s foundational flaw]] and [[https://​thegradient.pub/​how-to-fix-rl|How to fix reinforcement learning]]   * [[https://​thegradient.pub/​why-rl-is-flawed/​|Reinforcement learning’s foundational flaw]] and [[https://​thegradient.pub/​how-to-fix-rl|How to fix reinforcement learning]]
 +
 +==== Imitation Learning ====
 +
 +In Reinforcement Learning there was no expert to help learn a policy, Imitation Learning uses an expert who already has a good policy. ​ In Reinforcement Learning we care about learning the best policy, in Imitation Learning we mostly care about getting a good enough policy, if we want a much better one we find a better expert.
 +
 +  * [[https://​sites.google.com/​view/​icml2018-imitation-learning|Imitation Learning Tutorial from ICML 2018]]
 +
  
 =====THREAD4:​ Recent Achievements ===== =====THREAD4:​ Recent Achievements =====
Line 143: Line 149:
   * The [[wp>​Go_(game)|Go]] series [[https://​vk.com/​doc-44016343_437229031?​dl=56ce06e325d42fbc72|AlphaGo]],​ [[https://​deepmind.com/​research/​publications/​mastering-game-go-without-human-knowledge|AlphaGo Zero]] and [[https://​arxiv.org/​abs/​1712.01815|AlphaZero]]. ​   * The [[wp>​Go_(game)|Go]] series [[https://​vk.com/​doc-44016343_437229031?​dl=56ce06e325d42fbc72|AlphaGo]],​ [[https://​deepmind.com/​research/​publications/​mastering-game-go-without-human-knowledge|AlphaGo Zero]] and [[https://​arxiv.org/​abs/​1712.01815|AlphaZero]]. ​
   * [[https://​selfdrivingcars.mit.edu|Self driving cars]]   * [[https://​selfdrivingcars.mit.edu|Self driving cars]]
 +  * [[https://​www.youtube.com/​watch?​v=24AX4qJ7Tts|A Neural Network Model That Can Reason]]
  
 Interesting components: Interesting components:
Line 151: Line 158:
   * [[https://​twimlai.com/​twiml-talk-195-milestones-in-neural-natural-language-processing-with-sebastian-ruder/​|Milestones in Neural Natural Language Processing]] and [[https://​nlpprogress.com|current results]]   * [[https://​twimlai.com/​twiml-talk-195-milestones-in-neural-natural-language-processing-with-sebastian-ruder/​|Milestones in Neural Natural Language Processing]] and [[https://​nlpprogress.com|current results]]
   * [[https://​arxiv.org/​abs/​1708.07120|Super-Convergence:​ Very Fast Training of Neural Networks Using Large Learning Rates]]   * [[https://​arxiv.org/​abs/​1708.07120|Super-Convergence:​ Very Fast Training of Neural Networks Using Large Learning Rates]]
 +
  
 ==== Open Source Implementations ==== ==== Open Source Implementations ====
Line 180: Line 188:
   * [[http://​www.thetalkingmachines.com|Talking Machines]]   * [[http://​www.thetalkingmachines.com|Talking Machines]]
   * [[https://​twimlai.com|This Week in Machine Learning & AI ]]   * [[https://​twimlai.com|This Week in Machine Learning & AI ]]
- 
-People: 
-  * [[wp>​Geoffrey_Hinton|Geoff Hinton]] 
-  * [[wp>​Yoshua_Bengio]] [[http://​www.iro.umontreal.ca/​~bengioy/​yoshua_en/​|Uni Montreal]] 
-  * [[wp>​Yann LeCun]] [[https://​www.facebook.com/​yann.lecun|Facebook]] [[https://​twitter.com/​ylecun|Twitter]] 
  
 Organisations:​ Organisations:​
Line 196: Line 199:
  
 Conferences:​ Conferences:​
 +  * [[https://​icml.cc|International Conference on Machine Learning (ICML)]]
   * [[https://​neurips.cc|Neural Information Processing Systems (NeurIPS)]]   * [[https://​neurips.cc|Neural Information Processing Systems (NeurIPS)]]
   * [[https://​iclr.cc|International Conference on Learning Representations (ICLR)]]   * [[https://​iclr.cc|International Conference on Learning Representations (ICLR)]]
agi.1545060271.txt.gz · Last modified: 2018/12/17 15:24 by admin