Deep learning with CoW

    This site uses cookies. By continuing to browse this site, you are agreeing to our Cookie Policy.

    • Deep learning with CoW

      I feel that it is very wonderful that this CoW has also started using deep learning.
      However, the only example of application that has been published at present is automatic coloring of portrait images of weapons.
      So, I personally have both negative and positive interest in its development in same time.

      Deep learning is a new technology based on an uncertain hypothesis that the solution of various optimization problems “ may be abel to be “ emulated only by combining mechanical operations.

      Therefore operators will take an extremely high risk, but will gain a powerful tool in return.
      So, I want to propose to chat about future CoW and deep learning implications in this thread.

    • I'm sorry, pod, but I have no idea what you're talking about... but "coloring of portrait images of weapons" doesn't seem very useful? Does Bytro really use this Deep Learning for that??
      When the fake daddies are curtailed, we have failed. When their roller coaster tolerance is obliterated, their education funds are taken by Kazakhstani phishers, and their candy bars distributed between the Botswana youth gangs, we have succeeded.
      - BIG DADDY.

    • Yea, bit unclear. When I first read "deep learning", I was assuming you were talking about the behaviour of AI (sorta evolving over time kinda came to mind, but to be honest I've no clue what deep learning actually is). What's the auto colouring of images gotta to with this?
      Have an amazing rest of your day ^^

      "Everything is impermanent. The only thing that is permanent it impermanence itself."

      Need support? ---> Send a ticket here!

      dxter's CoW Battle Calculator ---> Use it here!

      :tumbleweed:

      o7

    • _Pyth0n_ wrote:

      Yea, bit unclear. When I first read "deep learning", I was assuming you were talking about the behaviour of AI (sorta evolving over time kinda came to mind, but to be honest I've no clue what deep learning actually is). What's the auto colouring of images gotta to with this?
      It's true that CoW hasn't made it clear though, I think it's almost certain that the image was transformed using a method called CNN.
      And CNN is a kind of deep learning.
      That was the such context. I'm sorry.

      K.Rokossovski wrote:

      I'm sorry, pod, but I have no idea what you're talking about... but "coloring of portrait images of weapons" doesn't seem very useful? Does Bytro really use this Deep Learning for that??
      And It's not that it's not useful though, I think that it may be too useful technology in a also bad way too.

      The post was edited 8 times, last by pod_than ().

    • In the country where I live, a TV series called “Pokemon” that a creation so famous that would probably have a hard time finding someone whose name don't know is still broadcasted.
      And if anyone refer to the page “Dennō Senshi Porygon” on Wikipedia, I think everyone can understand the gist, that images or portraits can injury humans too.
      Of course, about this one case or example, I think that photosensitive epilepsy is clearly recognized by people who are platformers though, it is virtually impossible to verify the risks that exist in a huge number.

    • _Pyth0n_ wrote:

      Yea, bit unclear. When I first read "deep learning", I was assuming you were talking about the behaviour of AI (sorta evolving over time kinda came to mind, but to be honest I've no clue what deep learning actually is). What's the auto colouring of images gotta to with this?
      I was thinking of that too directly. but then what does the Deep Learning/AI as a source, the majority of players, successful players, average players? It should learn from the right persons to become at least a somewhat intelligent approach to "AI", not building at all is at least an ignore criterium ;)

    • Gen. Smit wrote:

      _Pyth0n_ wrote:

      Yea, bit unclear. When I first read "deep learning", I was assuming you were talking about the behaviour of AI (sorta evolving over time kinda came to mind, but to be honest I've no clue what deep learning actually is). What's the auto colouring of images gotta to with this?
      I was thinking of that too directly. but then what does the Deep Learning/AI as a source, the majority of players, successful players, average players? It should learn from the right persons to become at least a somewhat intelligent approach to "AI", not building at all is at least an ignore criterium ;)
      It's too unclear my to understand the meaning properly though, if you translate it directly, is this what it means?
      ・No players have reported any health hazards in the past.
      ・The thread owner has never developed an AI, or even if there is, his teachers is not intelligent enough ( maybe ).
      ・Therefore, the point of the thread owner is not worth listening to. It is just one worthless opinion.

      The post was edited 2 times, last by pod_than ().

    • deep learning is a form of AI based on Art. Neural Networks, thats how I knew them, its often used in molecular biology to learn recognize (signal) patterns in biomolecules. So yes it could help AI to become better at analysis, choice of response and execution of the response. It should even be able to adapt with 1 hrs time frame , too much iteration could make it difficult for AI to decide or to battle AI (near perfect activity)

    • Gen. Smit wrote:

      deep learning is a form of AI based on Art. Neural Networks, thats how I knew them, its often used in molecular biology to learn recognize (signal) patterns in biomolecules. So yes it could help AI to become better at analysis, choice of response and execution of the response. It should even be able to adapt with 1 hrs time frame , too much iteration could make it difficult for AI to decide or to battle AI (near perfect activity)
      I see.
      Early stop is usually used to avoid over fitting, but it may be possible to reduce performance with some ingenuity.

    • pod_than wrote:

      In the country where I live, a TV series called “Charizard Pokemon” that a creation so famous that would probably have a hard time finding someone whose name don't know is still broadcasted.
      And if anyone refer to the page “Dennō Senshi Porygon” on Wikipedia, I think everyone can understand the gist, that images or portraits can injury humans too.
      Of course, about this one case or example, I think that photosensitive epilepsy is clearly recognized by people who are platformers though, it is virtually impossible to verify the risks that exist in a huge number.
      Almost everyone knows its name! And you're absolutely right about the episode "Dennō Senshi Porygon." It's a prime example of how images or flashing lights can potentially harm people, particularly those with photosensitive epilepsy. While the risks are evident to many of us, it's a real challenge to fully assess the dangers in a vast and diverse audience. It's a reminder of the importance of considering the well-being of all viewers when creating content.

      The post was edited 1 time, last by SmasherK ().