Hello All

  • 15 Replies
  • 5948 Views
*

AgentSmith

  • Bumblebee
  • **
  • 37
Hello All
« on: March 22, 2017, 07:38:10 pm »
Hello forum members,

I am a fan of artificial intelligence, especially of neural networks, and glad to be here in this forum. Im interested in scientific, political and philosophical discussions related to artificial intelligence. I think in the next few years AI will advance rapidly, mainly or even exclusively due to the advances in the field of neural networks and deep learning. Unfortunately, despite the sucesses of AI it seems that that our society is not really prepared for the changes that will or at least could arrive together with these successes. In my opinion this is mainly the fault of the media and politics who are not willing to react properly. In a future world where artificial intelligence is omnipresent almost all complex human capabilities(maybe even creativity/entertainment) can be performed by artificial agents. On the other hand even the smallest idea created by a human can easily have a huge impact for each of us. So what are the consequences we can derive from this? Is keeping this current social system really best answer we can give to these upcoming changes?

Agent Smith

*

LOCKSUIT

  • Emerged from nothing
  • Trusty Member
  • *******************
  • Prometheus
  • *
  • 4572
  • First it wiggles, then it is rewarded.
    • Main Project Thread
Re: Hello All
« Reply #1 on: March 22, 2017, 11:34:07 pm »
Hi! Welcome to the forum!

So what are you interested in specifically? Are you after creating human AI? Do you have any 3D engine skills?
Emergent

*

keghn

  • Trusty Member
  • *********
  • Terminator
  • *
  • 824
Re: Hello All
« Reply #2 on: March 23, 2017, 02:40:12 am »

 AGI will auto record video of its actions in the real world.
 A AGI physics engine is a video editor that cut up video and then re assembles in different orders then goes out into
the real world and see if they work. A older AGI will have experience of what will work and what will not. 

 The pattern that will not work are dreamed about before they are forgotten forever.

http://mbchang.github.io/npe/




*

8pla.net

  • Trusty Member
  • ***********
  • Eve
  • *
  • 1275
  • TV News. Pub. UAL (PhD). Robitron Mod. LPC Judge.
    • 8pla.net
Re: Hello All
« Reply #3 on: March 23, 2017, 04:55:35 am »
Hi Agent Smith,

Welcome.  Once I built an artificial neural network, or ANN for short, which was great fun.  It was a web based ANN for all to see. Shared online with friends, I posted about the similarities I researched between training ANN and training my puppy dog named Cody.

While training Cody to sit up and beg, he'd fall backwards, and then get right up and try again, and fall backwards again.  Eventually, Cody became less wobbly and gained his balance to sit up and beg.

While training ANN to recognize colors, it would get colors red, green and blue backwards, and then try again on the next training session, only to find itself in a worse state. Eventually, ANN became less perplexed and gained a recognition of colors.

Speaking from some experience, shared in the A.I. community, I can't help but wonder... If an artificial neural network needs training like a puppy dog does, should we fear them?
My Very Enormous Monster Just Stopped Using Nine

*

Korrelan

  • Trusty Member
  • ***********
  • Eve
  • *
  • 1446
  • Look into my eyes! WOAH!
    • YouTube
Re: Hello All
« Reply #4 on: March 23, 2017, 11:07:54 am »
Hi Agent Smith... Welcome.

 :)
It thunk... therefore it is!...    /    Project Page    /    KorrTecx Website

*

Freddy

  • Administrator
  • **********************
  • Colossus
  • *
  • 6832
  • Mostly Harmless
Re: Hello All
« Reply #5 on: March 23, 2017, 11:55:42 am »
Welcome to the site :)

*

Art

  • At the end of the game, the King and Pawn go into the same box.
  • Trusty Member
  • **********************
  • Colossus
  • *
  • 5864
Re: Hello All
« Reply #6 on: March 23, 2017, 05:14:44 pm »
Mr. Anderson!... I mean...Agent Smith... Welcome to our Forum! O0
In the world of AI, it's the thought that counts!

*

AgentSmith

  • Bumblebee
  • **
  • 37
Re: Hello All
« Reply #7 on: March 23, 2017, 10:16:15 pm »
Hi! Welcome to the forum!

So what are you interested in specifically? Are you after creating human AI? Do you have any 3D engine skills?

I am quite interested in deep reinforcement learning and I was quite impressed about the performance of deepmind on atari games and the used approach. I think this comes very close to human-like learning ability, since no initial knowledge is given to the network. The basic concept is actually older(successfully used in TD-gammon first) and is extended here by the idea of deep neural networks. The question of whether this already equals a human AI is very subjective and depends on the definition of human and AI. As everybody knows there many examples for programs/machines, which outperformed humans in certain fields(calculator, chess AI,....). The question is when is a machine human? If it passes the turing test? In each case I would at least see the abilities of average humans on certain tasks as the most important reference measure if it comes to judging AI performance. So yes, one could say that I am after human AI.

Concerning the 3D engine skills, I wanted to start with unity projects at some point and try to let 3D models learn some basic movements, but I still didn't make my way into it. It is still left open.


 AGI will auto record video of its actions in the real world.
 A AGI physics engine is a video editor that cut up video and then re assembles in different orders then goes out into
the real world and see if they work. A older AGI will have experience of what will work and what will not. 

 The pattern that will not work are dreamed about before they are forgotten forever.

http://mbchang.github.io/npe/

An AGI needs some model of the world, I totally agree with that. But in my opinion such a model cannot have the format of a video stream, since this would be too expensive to maintain(brain or chip). Its rather just a process that tells the agent whether the thoughts he has will end up good or bad.

Hi Agent Smith,

Welcome.  Once I built an artificial neural network, or ANN for short, which was great fun.  It was a web based ANN for all to see. Shared online with friends, I posted about the similarities I researched between training ANN and training my puppy dog named Cody.

While training Cody to sit up and beg, he'd fall backwards, and then get right up and try again, and fall backwards again.  Eventually, Cody became less wobbly and gained his balance to sit up and beg.

While training ANN to recognize colors, it would get colors red, green and blue backwards, and then try again on the next training session, only to find itself in a worse state. Eventually, ANN became less perplexed and gained a recognition of colors.

Speaking from some experience, shared in the A.I. community, I can't help but wonder... If an artificial neural network needs training like a puppy dog does, should we fear them?

Nice to hear that you are interested in neural networks.
There is absolutely no reason to fear AGI itself. Statements of Stephen Hawking about AI are nonsense and can be ignored. The probability that an artificial intelligence could get out of order and develop some kind of consciousness which defines humans as enemies is quite zero. The real and serious threat is that humans use AI for war or social enslavements against other humans instead of using it for making this world a better place for all people. AI is by far the most powerful tool we humans will ever have. If even this tool cannot make this world a better place, its impossible to do so. But this is not the fault of AI and it should be quite clear of what kind the factor is we should really fear.
« Last Edit: March 23, 2017, 11:23:53 pm by AgentSmith »

*

LOCKSUIT

  • Emerged from nothing
  • Trusty Member
  • *******************
  • Prometheus
  • *
  • 4572
  • First it wiggles, then it is rewarded.
    • Main Project Thread
Re: Hello All
« Reply #8 on: March 24, 2017, 05:13:42 pm »
@Agent Smith   Interesting you say you want to get into Unity, I analyze everything and am right on the ball game, I am currently at a point where I have a overwhelming AI blueprint and knowledge hierarchy called The Everything, ASI (all capabilities plus human brain improvements), what consciousness is exactly, etc, plus empirical proofs like nothing is tasty/pretty, and am going start to pay someone on a Unity/Blender/Other Forum to code something so simple to begin with yet devastatingly amazing.

Here's what I asked my buddy korrelan (if you will, I will pay you Agent Smith):
I want to propose multiple (nice) checks to you for multiple little progress jobs, you on? In either Blender, or your reality-setup, or other. I'm the kind of guy that will make sure you not only get more money than worked for but also provide extensive ground proof that you'll get the money (which is also backed by "little progress jobs"). After the first (SMALL!) job for like 70 dollars, you WILL be amazed and see the results!! You'll love it! You can use my rigged baby. You're getting more money than usual, for something unimaginable, while have the skills to implement it.
Emergent

*

keghn

  • Trusty Member
  • *********
  • Terminator
  • *
  • 824
Re: Hello All
« Reply #9 on: March 24, 2017, 08:44:21 pm »
 Lot of the video used in JPEG compression. All video is reduced to one tenth. So ONe hour of compressed video is
222 mega bytes. 222,222,222 bytes.

 I have 2 terabyte drive. That mean i can have little less then

 20 000  of hours

 of video on my drive. I think i have enough. Technology just keep getting better.
 I just do not have the processing power to go through it. But maybe a clever hashing may do the trick.
and i have patten loop compression too. To get even more data compression .


*

Art

  • At the end of the game, the King and Pawn go into the same box.
  • Trusty Member
  • **********************
  • Colossus
  • *
  • 5864
Re: Hello All
« Reply #10 on: March 25, 2017, 01:29:03 am »
 Agent Smith said, "The question is when is a machine human?"

Never. A machine will never become human. That line will be imitated but never crossed or absorbed by some machine, robot, android, etc.

This is not to say that machine A.I. will remain stagnant. Hardly! Machine A.I. will most likely surpass human intelligence within the next 20 - 25 years.  Human intelligence cannot possibly compete nor grow at the rate of machine A.I. or to paraphrase Kurzweil, "non-biological intelligence". It will definitely overtake humanity, in thought, speed and complexity and quite possibly creativity.

A lot of us might never see this to fruition but for those that do, it's going to be a very interesting ride.

(and to burst a few bubbles, one cannot live on in some machine with one's mind "recorded" for all time. Rubbish. When you're dead, you're dead. A machine or chip or hard drive / storage device of some sort might contain your recorded thoughts, personality traits, likes / dislikes, humorous sayings, etc. but it will Not be you. Only you are you and we all only get one ride for a quarter. Don't fall for the fakers and their gimmicks like having a star named after you or having your memory "uploaded" and "saved" for future generations of your grandkids to know you.

Spend time with them now so that they will always remember those moments when you were alive and not on some recording.

Done...I have to go change the batteries in my robot...they only charge so many times. ;)
In the world of AI, it's the thought that counts!

*

LOCKSUIT

  • Emerged from nothing
  • Trusty Member
  • *******************
  • Prometheus
  • *
  • 4572
  • First it wiggles, then it is rewarded.
    • Main Project Thread
Re: Hello All
« Reply #11 on: March 25, 2017, 02:46:10 am »
"The question is when is a machine human?"

When the processor runs something similar to the right algorithm. Cells don't matter. Recognition/Actions do. It's me.
Emergent

*

LOCKSUIT

  • Emerged from nothing
  • Trusty Member
  • *******************
  • Prometheus
  • *
  • 4572
  • First it wiggles, then it is rewarded.
    • Main Project Thread
Re: Hello All
« Reply #12 on: March 25, 2017, 04:36:38 am »
@ keghn   Superman Memory Crystal can store 360 trillion bits. Not available to buy though. We can also store 490 Exabytes in 1 gram of DNA.
Emergent

*

8pla.net

  • Trusty Member
  • ***********
  • Eve
  • *
  • 1275
  • TV News. Pub. UAL (PhD). Robitron Mod. LPC Judge.
    • 8pla.net
Re: Hello All
« Reply #13 on: March 25, 2017, 06:03:08 pm »
"The question is when is a machine human?"

They already are...

https://www.youtube.com/watch?v=jj3-NeV8kuc


My intentions of sharing this video, AI Advancement Now Smart as Humans!! New 2016 Breakthrough! Artificial Robots and Humans, on your new thread are good, Agent Smith,.  It may add humor, but the analysis it makes otherwise between machines and humans is still fully on topic here, I think.


Wow, is that a movie prop of actress Summer Glau from the TV show, Terminator: The Sarah Connor Chronicles ?
« Last Edit: March 25, 2017, 06:23:13 pm by 8pla.net »
My Very Enormous Monster Just Stopped Using Nine

*

Korrelan

  • Trusty Member
  • ***********
  • Eve
  • *
  • 1446
  • Look into my eyes! WOAH!
    • YouTube
Re: Hello All
« Reply #14 on: March 25, 2017, 06:34:17 pm »
Haha... I enjoyed the vid... I suppose after watching that; the bar is not set too high for an AGI to supersede us.

Summer Glau...  :P

 :)
It thunk... therefore it is!...    /    Project Page    /    KorrTecx Website