Autres_Dimensions_4_CNN.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"Autres_Dimensions_4_CNN.ipynb","provenance":[{"file_id":"1UZc6OSSZYCQeTw_Znbg1CLB3q5FThDck","timestamp":1582650080190},{"file_id":"1gJUxZ7Clh1Hj-_PLcJyKLsdeZAkrED_O","timestamp":1578408594414}],"collapsed_sections":[]},"kernelspec":{"name":"python3","display_name":"Python 3"}},"cells":[{"cell_type":"markdown","metadata":{"id":"37LNuwykzt3X","colab_type":"text"},"source":["# **Les aventuriers des Axes Perdus**\n","#Epoch 4\n","\n","***Olivier Thöni*** *(novice en IA, et sinon... professeur de mathématiques)*\n","\n","*School of AI - Le Mans*\n","\n","Lien vers le présent NoteBook : https://frama.link/OT_dimension4    \n","\n","\n","\n"]},{"cell_type":"markdown","metadata":{"id":"EigAUqpVFdPl","colab_type":"text"},"source":["# **1. Diminuer la dimensionnalité**\n","\n","## **1.1.la régression linéaire**\n","\n","https://frama.link/OT_dimensions \n"]},{"cell_type":"markdown","metadata":{"id":"4U1AMWGaGAhW","colab_type":"text"},"source":["## **1.2. L' \"Analyse en Composantes Principales\" (ACP)**\n","\n","### **1.2.a/ la classification des iris**\n","\n","https://frama.link/OT_dimension2  \n","\n"]},{"cell_type":"markdown","metadata":{"id":"gfBrGD5QGm4-","colab_type":"text"},"source":["### **1.2.b/ Les décathloniens**\n","\n","\n","https://www.kaggle.com/capedcrusader/pca-on-a-random-dataset/notebook (en R)\n","\n","\n","### **1.2.c/ Un usage d'une ACP simplifiée : la SVD pour la compression d'image**\n","\n","https://frama.link/OT_dimension3 \n","\n","\n","\n","\n"]},{"cell_type":"markdown","metadata":{"id":"_PSHUhZKIb6x","colab_type":"text"},"source":["## **1.3. \"Réseau de Neurones Convolutionnel\" (CNN)**\n","\n","\n","![CNN](https://drive.google.com/uc?id=1Zl_mA_CL1lONVVgGr4ssryBYdvZVHEMj)\n","\n","Le réseau de convolution (CNN) pour la classification d'images : en passant dans un réseau de neurones convolutionnel, chaque \"tuilage\" revient à faire diminuer la **taille** des matrices en présence.\n","\n","\n","http://scs.ryerson.ca/~aharley/vis/conv/ \n","\n","![CNN](https://drive.google.com/uc?id=1N889VtPjTwgIeZeg9J-EHYyPnVKgjLAR)\n"]},{"cell_type":"markdown","metadata":{"id":"Qr4MIOqNJFre","colab_type":"text"},"source":["# **2. Récupérer l'information perdue : \"AutoEncoder\"**\n","\n","https://frama.link/OT_dimension5 "]}]}