[mlpack] Google Summer of Code

[Marcus Edel]

Hello Papoudakis,

thanks for getting in touch and welcome.

> I have a strong background in deep reinforcement learning since my diploma
> thesis is Deep Reinforcement Learning Agent Collaboration for the Game of Doom.
> Doom is a 3D fps game which is more challenging than atari games. I have already
> implemented asynchronous advanced actor-critic algorithm using tensorflow [1].
> Additionally you can check my blog in which I post articles about deep
> reinforcement learning [2].

Your diploma thesis sounds really nice would be interesting to take a look at,
but I guess it's not in English? Also, I agree solving a level of Doom is is
more challenging than solving a level of an atari game. The nice thing is since
we can use OpenAI's Gym toolkit we are able to experiment with a lot of
different games including Doom and even GTA if we like to.

> Finally, I have experience using C++ pthreads which are essential for
> implementing asynchronous algorithms.

Note we don't use pthreads rather we use the OpenMP directives which allow's us
to parallelize code with minimal extra code. It's platform independent and the
code looks a lot cleaner.

> I will start experimenting with mlpack code after 20th of March because right
> now  I have to present my thesis. My proposal for google summer of code will
> probably be implementing a3c and n-steps Q-learning algorithm.

Sounds good, we can brainstorm details after 20th of March if you like, best of
luck with your thesis.

> Excuse me if you find the following question disturbing. I have to decide if I
> will graduate or not. Do you think that I have good chances to trust we with
> this project? I must be a student in order to apply for google summer of code.

We have no idea how many people are going to get in touch, so I can't answer
that.

I hope this is helpful, let us know if you have any more questions.

Thanks,
Marcus

> On 1 Mar 2017, at 11:28, Papoudakis Giorgos <giwrpapoud at gmail.com> wrote:
> 
> Dear Marcus,
> I am Papoudakis Giorgos and I am a student in the department of Electrical and Computer Engineering at Aristotle University of Thessaloniki in Greece. I would like to apply for google summer of code in mlpack for implementing reinforcement learning algorithms. I have a strong background in deep reinforcement learning since my diploma thesis is Deep Reinforcement Learning Agent Collaboration for the Game of Doom. Doom is a 3D fps game which is more challenging than atari games. I have already implemented asynchronous advanced actor-critic algorithm using tensorflow [1]. Additionally you can check my blog in which I post articles about deep reinforcement learning [2]. Finally, I have experience using C++ pthreads which are essential for implementing asynchronous algorithms.
> 
>  I will start experimenting with mlpack code after 20th of March because right now  I have to present my thesis. My proposal for google summer of code will probably be implementing a3c and n-steps Q-learning algorithm. 
> 
> Excuse me if you find the following question disturbing. I have to decide if I will graduate or not. Do you think that I have good chances to trust we with this project? I must be a student in order to apply for google summer of code.
> 
> Thank you very much time. I am looking forward to hearing from you in due course.
> 
> Sincerely Yours,
> Papoudakis Giorgos.
> 
> [1] https://github.com/papoudakis/Asynchronous_RL <https://github.com/papoudakis/Asynchronous_RL>
> [2] https://papoudakis.github.io/ <https://papoudakis.github.io/>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://knife.lugatgt.org/pipermail/mlpack/attachments/20170301/9ee8c701/attachment.html>