Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Thanh-Binh
    @Thanh-Binh
    I mean EOgmaDrive
    222464
    @222464
    @Thanh-Binh what do you mean? It is using video
    it has a simulated camera
    Thanh-Binh
    @Thanh-Binh
    @222464 cool. I like it. Can I test it?
    222464
    @222464
    @Thanh-Binh sure, it uses python though
    to run it, you need pyaogmaneo, opencv, and of course Donkey-Gym
    Thanh-Binh
    @Thanh-Binh
    @222464 thanks
    Thanh-Binh
    @Thanh-Binh
    @222464 in the mean time i found a c++ version of lunar lander
    Do you think that we can use it ?
    222464
    @222464
    It says in the readme that it is not complete yet @Thanh-Binh
    also, it uses a modified box2d. I do not understand why though - the original python version didn't modify box2d as far as I know
    Thanh-Binh
    @Thanh-Binh
    Are you aware of this repository:
    Some environments of OpenAI Gym are implemented in C++
    222464
    @222464
    @Thanh-Binh No, but like the other one it doesn't seem to have graphics yet
    most stuff is in Python unfortunately
    Thanh-Binh
    @Thanh-Binh
    @222464 OK thanks
    222464
    @222464
    @Thanh-Binh new robot video with AOgmaNeo: https://www.youtube.com/watch?v=Dw6Upzcom0A
    Thanh-Binh
    @Thanh-Binh
    Cool! Congratulation!
    Does it include training phase?
    222464
    @222464
    @Thanh-Binh it is pre-trained to the hardcoded policy, then trained by walking around in real life for 2 minutes or so
    I can try training it for longer, but the motors get really hot on this robot
    Thanh-Binh
    @Thanh-Binh
    However, it works very well!
    super!
    222464
    @222464
    @Thanh-Binh new video recall demo (with audio): https://www.youtube.com/watch?v=64xac-fITd0&feature=youtu.be
    jkyle1234
    @jkyle1234
    Hey Eric saw the video and looks interesting. I have a few questions about it. When we say recall what exactly is it recalling and is this something stored in the weights or offline in a file or what not? Was also wondering what the difference between AOgmaNeo vs OgmaNeo2? Is AOgmaNeo a different architecture or just stripped down to run on a micro-controller? Thanks.
    222464
    @222464
    @jkyle1234 It doesn't use anything that isn't stored in the weights or activations of the network. For AOgmaNeo vs OgmaNeo2, AOgmaNeo started as just a stripped-down version of OgmaNeo2, but it is the best API for desktop as well now. It's basically better in every way!
    Thanh-Binh
    @Thanh-Binh
    @222464 i have just looked at the current version, it has some new encoder. Can you share your idea? Thanks
    222464
    @222464
    @Thanh-Binh yes, it uses a "dual encoder" now - it's really 2 encoders with different strengths (one is generative and the other discriminative). The generative one is the older ESR encoder, and the new discriminative one uses prediction errors to update its representation. Individually they are not good on all tasks, but together they complement each other and provide superior performance.
    it can now learn to add numbers with a single layer, and still perform long time horizon predictions
    It also seems to help the reinforcement learning as well, better performance all around
    by itself, the older ESR encoder pooled different inputs pretty well but wasn't able to discriminate certain inputs. The error-driven encoder could discriminate really well but could not pool inputs together well, resulting in a very noisy representation. Together they augment each other and it gets the best of both
    Since it is 2 encoders now, it is slightly slower than before. But I think it is worth it
    Thanh-Binh
    @Thanh-Binh
    @222464 thanks for your explain!
    Adam
    @alienatorZ
    Will you be doing any updates to the white paper or presentation?
    222464
    @222464
    @alienatorZ yes, there is a new guide that will soon be complete
    Adam
    @alienatorZ
    Great!
    Thanh-Binh
    @Thanh-Binh
    @222464 is the new guide available?
    could you pls update a C++ example?
    thanks
    222464
    @222464
    @Thanh-Binh Guide is now up! https://github.com/ogmacorp/AOgmaNeo/blob/master/AOgmaNeo_User_Guide.pdf It doesn't include code examples, it just describes how the system works briefly, so that one knows what the different concepts are. I will update the C++ and Python code examples next to show how to use the library (C++ first, since python still has working examples)
    222464
    @222464
    @Thanh-Binh C++ demos have been updated to run, but are not well documented yet: https://github.com/ogmacorp/OgmaNeoDemos/tree/aogmaneo
    Thanh-Binh
    @Thanh-Binh
    @222464 super & thanks. I will take a look soon..
    Thanh-Binh
    @Thanh-Binh
    @222464 i tested it with your demo and found that the perform for wavy and car racing app are worse than the previous version. Maybe the parameter set need to be optimized..
    222464
    @222464
    @Thanh-Binh Hmm, it seems to work better than the old version for me - but maybe I am mistaken
    It does both of those pretty easily
    Unless you mean runtime speed, in which case yes that is expected, it is slightly slower