Slides: V0.5 intro + skeleton

slides/01-intro.marp.md → slides/00-intro.md

@@ -178,14 +178,94 @@ _color: black
 
 ---
 ## C'est quoi apprendre ?
+<!-- 
+
+Behaviorisme: apprendre = savoir _faire_ 
+Acquis vs Inné : danse des abeilles vs culture des chimpanzés
+Conditionnement opérant (Pavlov et sa cloche)
+
+-->
 ---
-Le par-coeur, un problème ? 
+
+## C'est quoi apprendre ?
+#### voir...
+- [WP: Learning](https://en.wikipedia.org/wiki/Learning)
+- [WP: _Machine_ Learning](https://en.wikipedia.org/wiki/Machine_Learning)
+- [WP: Operant Conditioning](https://en.wikipedia.org/wiki/Operant_conditioning)
 
 ---
-Guessing the teachers password 
+### Conditionnement 
+## Opérant
+![bg right 99%](./img/01-operant.png)
+
+---
+<!-- 
+_color: black
+_backgroundColor: white 
+-->
+### Reinforcement
+##  Learning
+
+![bg right 90%](./img/01-reinforcement.png)
 
 ---
-Learning to forget
+### Le par-coeur, un problème ? 
+
+---
+### Guessing the teachers password 
+<!-- 
+Proper beliefs pay rent by functioning as anticipation-controllers; 
+a proper belief is not just a string of words, but a cognitive structure fully integrated with the believer's model of how the world simply is. 
+Unfortunately, many students in traditional schools become accustomed to guessing the teacher's password, repeating back words that make it superficially sound as if they understand the topic, when in fact they don't.
+ -->
+
+<br />
+<br />
+<br />
+
+- Voir [LessWrong](https://www.lesswrong.com/posts/NMoLJuDJEms7Ku9XS/guessing-the-teacher-s-password)
+---
+## Apprendre 
+## c'est oublier
+
+
+![bg right 90%](./img/01-information.png)
+
+<!-- 
+Initially,weights are randomly initialised. 
+So barely anything is known about the correct output. 
+
+With successive layers the mutual information about input decreases and the information in hidden layers about the output is low as well.
+ -->
+---
+## Apprendre 
+## c'est oublier
+
+![bg right 90%](./img/01-information2.png)
+
+<!-- 
+As we train the Neural Network the plots start moving up, signifying gain of information about the output.
+
+But.
+
+Plots also start shifting towards the right side, signifying increase of information in latter layers about the input.
+ -->
+---
+## Apprendre 
+### c'est 
+###### (pas tout) 
+#### oublier
+
+
+- ###### [[Towards Data Science] _Information theory of Neural Networks_](https://towardsdatascience.com/information-theory-of-neural-networks-ad4053f8e177)
+![bg right 90%](./img/01-information3.png)
+
+<!-- 
+
+After the compression phase, the plots starts shifting towards the top but also to left side.
+
+This signifies, with successive layers, there is loss of information about the input and what’s retained in the last layer is the lowest entropy information about the output.
+-->
 
 ---
 Comment je sais ce que je sais pas ? 
@@ -197,13 +277,19 @@ Comment je sais ce que je sais pas ?
 MIT COVID FAILS 
 
 ---
-Gaming the game: IAs flemmardes
+### Gaming the game: IAs flemmardes
+
+<br />
+
+- [Régalez vous <br /> avec ce tableau](https://docs.google.com/spreadsheets/u/1/d/e/2PACX-1vRPiprOaC3HsCf5Tuum8bRfzYUiKLRqJmbOoC-32JorNdfyTiRRsR7Ea5eWtvsWzuxo8bjOxCG84dAg/pubhtml)
+
+![bg right 90%](./img/01-cheater.png)
 
 ---
-Short-term goals vs long-term goals 
+### Short-term goals vs long-term goals 
 
 ---
-Dangerous goals : Paperclip maximizer? 
+### Dangerous goals : Paperclip maximizer? 
 
 
 ---
@@ -267,11 +353,3 @@ Supervisé ou non
 
 ---
 ### Yudkowsky 
\ No newline at end of file
----
-## Adversarial perturbations 
-
-
-## Méthodes 
-### Train/test/val split
-
-## Outils
\ No newline at end of file

slides/01-choisir.md

+---
+marp: true
+theme: uncover
+color: #eee
+colorSecondary: #333
+backgroundColor: #111
+paginate: true
+---
+
+# Choisir un Modèle
\ No newline at end of file

slides/02-entrainer.md

+---
+marp: true
+theme: uncover
+color: #eee
+colorSecondary: #333
+backgroundColor: #111
+paginate: true
+---
+
+# Entrainer son Modèle
\ No newline at end of file

slides/03-tester.md

+---
+marp: true
+theme: uncover
+color: #eee
+colorSecondary: #333
+backgroundColor: #111
+paginate: true
+---
+
+# Tester son Modèle
+
+
+---
+## Adversarial perturbations 
+

slides/04-utiliser.md

+---
+marp: true
+theme: uncover
+color: #eee
+colorSecondary: #333
+backgroundColor: #111
+paginate: true
+---
+
+# Utiliser un Modèle
+
+--- 
+
+HuggingFace!
+
+
+--- 
+
+## Outils
+
+Dev, Collab, etc
\ No newline at end of file