https://github.com/srush/aima-arguments
https://github.com/srush/aima-arguments
Last synced: about 1 year ago
JSON representation
- Host: GitHub
- URL: https://github.com/srush/aima-arguments
- Owner: srush
- Created: 2021-10-11T00:24:43.000Z (over 4 years ago)
- Default Branch: main
- Last Pushed: 2021-10-11T01:05:53.000Z (over 4 years ago)
- Last Synced: 2025-01-28T05:41:21.704Z (over 1 year ago)
- Size: 6.84 KB
- Stars: 7
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# AI: A Modern Approach?
In a fit of frustration, I wrote a troll post about AI: A Modern Approach.
https://twitter.com/srush_nlp/status/1446469104894873613
It led to a very interesting set of reactions. Here is a brief bibliography of the discussions that ensued:
### Camp 0 -> AI is exactly what is in R&N
* Lots of people feel this way!
### Camp 1 -> This confusion is really about the sociology of the field circa 1992
* Scruffy vs Neat : https://en.wikipedia.org/wiki/Neats_and_scruffies, https://monoskop.org/images/1/1e/McCorduck_Pamela_Machines_Who_Think_2nd_ed.pdf
* Symbolic vs Connectionist (Minsky): https://web.media.mit.edu/~minsky/papers/SymbolicVs.Connectionist.html
### Camp 2 -> AI is inherently about Agency
* Rationality of Herbert Simon: https://www.scielo.br/j/rep/a/CWfwPPVWKvLrndfxR9vYFHL/?lang=en
### Camp 3 -> AI is inherently about sequential decision making (Read a formal RL text)
* Algorithms for Decision Making: https://algorithmsbook.com/files/dm.pdf (this book looks amazing)
* Algorithms for Reinforcement Learning: https://sites.ualberta.ca/~szepesva/rlbook.html
### Camp 4 -> High-level is good, but R&N is not
* ML from Tom Mitchell's text: http://www.cs.cmu.edu/~tom/mlbook-chapter-slides.html
### Camp 5 -> ML is good, but prob ML texts are too informal
* Foundations of ML: https://cs.nyu.edu/~mohri/mlbook/
* Elements of Statistical Learning: https://web.stanford.edu/~hastie/ElemStatLearn/
* Understanding Machine Learning: https://www.cs.huji.ac.il/~shais/UnderstandingMachineLearning/
### Camp 6 -> PRML is on the right track, but doesn't have enough new stuff
* Prob ML - https://probml.github.io/pml-book/book1.html
### Camp 7 -> Philosophy is critical for AI, but R&N is too scattered
* Cambridge Handbook of AI: http://www.cambridge.org/9780521871426
### Camp X -> Arguments about Rigor are really important and we should have more of them
* Ali Rahimi's Test of Time Talk: https://www.youtube.com/watch?v=Qi1Yry33TQE