Lab Reinforcement learning

Cart Pole This video explains the problem we try to solve using reinforcement learning. On this page we will try to resume the lab we made in class and which parameter affect the efficiency of our model. QNetwork: In this section we define our network which is a fully connected leyer. Results: it returns the value function at the current for the action 0 on the left and the action 1 on the right 10 times because there are 10 states (we take the max and the agent choose the value)
Read more →

lab: 3D reconstruction

The goal of this lab was to implement a 3D reconstruction from multi-views of human pose. We had to reconstruct the 3D skeleton of a person from the matching joints of multi-views. To do so we had data: videos of subjects doing different movements OP2DTXT, containing text files of 25 joints on each sequences of each subject, this is the format of every line X and y are horizontal and vertical pxiels and r is the reliability score
Read more →

Labs on blender

Blender In the different labs we learn how to create different type of shapes thanks to the exctract tools. We also learnt how to change materials and their colours. Thanks to these labs I did these 3 3D creations : a launcher, my clio 1 car and my character that looks like a stonks version of Spirou. launcher I won’t go deep in this lab since everyone has the same object at the end, but it’s pretty cool so here is my launcher :
Read more →

Lecture 1

Affordances : Good example : gear level : Just by looking at the scheme on the gear level we instantly know how to use it. In fact, when the car is in neutral gear, just by looking at the scheme you understand that you have to put the gear level in the upper-left corner, and it’s the same when you want to shift gears even if we can’t see the mechanical pattern.
Read more →

Lecture 2

Lecture 4 : Let me introduce you Jonathan Grudin who is an American HCI researcher, he works as a principal design researcher at Microsoft and he is also a professor in the University of Washington Information School. He became famous since he received an award from the CHI Academy which is a group of people specialized in Human-Computer Interaction and since he is a pioneer of this field. As he worked on CSCW (Computer Supported Cooperative Work) he raised something that we will call later the “Grudin Problem” which explain why CSCW application fail and this is what we will deal with.
Read more →