This project is a fork of the https://github.com/nbatfai/SamuMovie repository, and it is used for an exercise of the High level programming languages course.

I've created a video converter, which transforms the given input video into 2D arrays of bool, and I've modified the original Samu to be able to learn from this data.

Example video

Samu (Nahshon) has captured a movie. This is an example of the manuscript entitled "Samu in his prenatal development".

This project uses the same COP-based Q-learning engine as Samu (Nahshon) to predict the next sentences of a conversation.

After the basic work on the joining of Conway's Game of Life and Samu's COP engine is done, now I have started to work on creating a similar but much simpler example in which Samu must learn the frames of the moving picture.

git clone https://github.com/nbatfai/SamuMovie.git
cd SamuMovie/
~/Qt/5.5/gcc_64/bin/qmake SamuLife.pro
make
./SamuMovie

https://youtu.be/XOPORbI1hz4

It is important to note that in the previous version the COP-based Q-learning has become trivial because after a short starting period it chooses the Q-action that was passed in as the actual cell state argument. It means that the agent does not predict the future but the present. This version has already been improved. It really predicts the future. It can be verify in SamuLife

https://youtu.be/yOZj6j1kVRg

Samu (Nahshon) http://arxiv.org/abs/1511.02889, https://github.com/nbatfai/nahshon

SamuLife https://github.com/nbatfai/SamuLife, https://youtu.be/b60m__3I-UM

SamuMovie https://github.com/nbatfai/SamuMovie, https://youtu.be/XOPORbI1hz4

SamuStroop https://github.com/nbatfai/SamuStroop, https://youtu.be/6elIla_bIrw, https://youtu.be/VujHHeYuzIk

SamuBrain https://github.com/nbatfai/SamuBrain

SamuCopy https://github.com/nbatfai/SamuCopy

SamuTicker https://github.com/nbatfai/SamuTicker

SamuVocab https://github.com/nbatfai/SamuVocab