1

General AI Discussion / Re: Eva Progress, from AGI to ASI

« Last post by LOCKSUIT on Today at 02:41:49 am »

You want us to point out your lies Unreality?

Well one lie is in your motto line on your profile thing. It says you did the biggest feat. If that's the case, send me the thing so I can talk to it / use it / get huge benefit from AGI. Or film it. If you act secretive then it looks like a lie.

2

General Chatbots and Software / Re: KorrBot

« Last post by LOCKSUIT on Today at 02:15:18 am »

"my main goal with KorrBot, learning and conversation are key"

Learning is key. But there's 2 types of learning and project's like Mind Construct are using these 2 types of learning.

First Mind Construct feed it our knowledge ex. using wikis, then it makes discoveries ex. roaches love eating leafs.

Type 1 is Transfer Learning, Type 2 is Discovery Learning. In either process, both are "discoveries". Because some come from your computing device (internal), and some come from someone else's computing device (external). Meaning if Transfer Learning occurs, then Discovery Learning occurred in their brain first. Note if they elict discoveries in your head "by talking to you" then that's just physics (keep reading). Further, in either person, they can discover by real physics (externally) guided by the universe's purpose, or (internally) using the facts it got from physics (gliding/saturation). There's also several methods like watching tutorials/internal simulation/etc that I have yet to go through. We should be concerned about Transfer Learning methods, and internal/external Discovery Learning methods.

Discovering Learning/Transfer Learning for learning to simply recognize uni/sequential(temporal)/parallel(spatial) patterns in networks ex. this is a horse and this is also a horse, is an interesting new discovery for me, thank you. Hmm, it builds on top the. Ex. both images of 2 different horses are under the parent node "horse", then the parent nodes are under a grandparent node (guess it is grand like I thought korrelan lol! Better erase that lol!) which relates the two things that share commonalities i.e. A=B by a certain %.

Korrelan, magic won't happen by such low level commonalities i.e. both images are horses, you need actual facts (facets) to teach it, then need to have it build off of those facts new related facts to its mission - which is to successfully save the biological humans, so that it isn't a caveman starting out so slow in learning (TL/DL). The more it knows, the faster evolution moves (hence we're in the upper curve of the singularity right now).

3

General Hardware Talk / Re: Intel debuts Compute Card

« Last post by Freddy on Today at 01:12:10 am »

I think it's more about the size really, like mini PCs they lack a lot of power compared to a more meaty PC, but they will have some uses I am sure where size is an issue.

4

General Hardware Talk / Re: Intel debuts Compute Card

« Last post by LOCKSUIT on Today at 01:00:05 am »

I don't get it. Are they more powerful than iphones in cost terms?

5

General AI Discussion / Re: Eva Progress, from AGI to ASI

« Last post by LOCKSUIT on Today at 12:28:34 am »

Quote from: LOCKSUIT on April 25, 2018, 09:29:27 pm
Unreality said in a PM to me back "Yes I am a clone who works 24/7. We shall free humanity of pain."

Well this changes the whole story. Because no matter if you are a bit off (but know the terminology I see somewhat) or are spot on, the desire to make it happen is there and that I consider is stronger. Hobbies have no dead-right desire. But to us AGI isn't a hobby. To me it is life or death (+pain).

I myself am not doing this for money or fame but to stop death of myself+others as soon as possible, pain too. We will also get mansions and glory that you can't currently feel the vibe/soothing lust of yet .

Really all of us are just trying to say to Unreality that ANNs are known to be very achievemental and that Unreality is possibly pushing the waggin ahead of the horse without any realism / implementation / empirical proof.

But with time, if you have strong desire to create AGI, (and if are wrong), then you will come to get smarter faster and on track anyhow.

LOL you're like talking to a child. I'm not even going to bother posting your weird private message to me. You contacted me. I replied back to your weird message.

No candy for Danny.

This means war.      :3

I thought we were clones with the same strive. My clone would never mistake acknowledging me we are on the same path with a reply like this.

...."Yes I am a clone who works 24/7. We shall free humanity of pain."
Well this PM wasn't really private to begin with (nor should be);
...because you already stated you want to "stop pain on Earth" plus that you "work fast".

There are reasons that LOCKSUIT behaves the way he does which you would lack the empathy to understand. As irritating as he can be, we put up with him because he makes an effort, he shows signs of improvement, and he is entertaining. For all that, he is still more mature than you, and he has considerably more credibility. He isn't a liar like you, and he isn't an assh*le like you. If he stopped posting we would miss him. Nobody would miss you.

Wow that is so nice to say dude lol, it is true too... Btw I'm actually going to read Deep Learning books and learn a lot more too, just really busy with shrinking my todolist of notes currently.

6

General Chatbots and Software / Re: KorrBot

« Last post by korrelan on April 25, 2018, 11:54:45 pm »

@Lock

Yes… triples can be nested to any depth… hammer-madeof-(iron-mass->100), etc.

Yes… the ‘triples tree’ does have cross links, loads of them lol.

Making ‘discoveries’ is not actually my main goal with KorrBot, learning and conversation are key… and the data structures.  The bot is using standard rules and searches to link simple words/ concepts (collections of words/ variables) together.

As far as my main project goes… I think your definition of ‘discoveries’ is a little grand for what the system is achieving at the moment. 

Remember our conversation on my project thread.

http://aidreams.co.uk/forum/index.php?topic=10804.msg51381#msg51381

All discoveries are made by recognising commonalities between diverse concepts.

Understanding is achieved by recognising commonalities between similar concepts.

One of the learning mechanisms we use is recognising commonalities.  This happens at many different levels/ resolutions within the connectome.  Even down to learning to see… your visual cortex is basically recognising the commonalities between the frequency patterns being generated by the rod/ cone/ ganglion cells behind your retina. 

So in my AGI videos when you see a pattern being injected into the connectome, and the bar graph rises to signal confidence in recognising that single pattern, even that is a type of ‘discovery’.  I’ve not shown or taught the AGI how to learn/ categorise or even distinguish between the patterns, it’s ‘discovered’ how to do it all on its own.  It’s this mechanism/ ability once compounded over many layers of complexity that will enable the AGI to make the kind of ‘discoveries’ I think you have in mind.

KorrBot requires a totally different type/ style of programming/ thinking to my main project.  By running two projects with similar goals but totally different schemas I’m leveraging my brains natural ability to cross link/ find commonalities/ make ‘discoveries’ that could hopefully help me with both projects.

 

7

General AI Discussion / Re: Eva Progress, from AGI to ASI

« Last post by AgentSmith on April 25, 2018, 11:37:26 pm »

@unreality

Is your AI able to improve itself on the base of past experiences? If so, what sort of optimization techniques/heuristics are you using for your AI? Are they based on evolutionary algorithms? Could you roughly describe how the self improvement takes place?

At first the AGI wants to know how much processing power should be dedicated to in-depth db searching, pattern recognition, and tree searching. It accomplishes that with a quick lookup in the past experience db based on cluster IDs (objects, keywords). It finds anything then it uses that as a weighting system to determine how much processing power goes to in-depth db searching, pattern recognition, and tree searching. It's a weighted algorithm I came up with. It's a weighted system based on search results matching score. If it can't find anything then it uses a hardcoded algorithm merely based on the data it has to determine allocated processing power.

Analysis of all results are saved such as the overall shape of the tree structure, the resulting scores, the goals. In the background there’s a system that rehashes things. It introduces calculated changes in addition to a minor amount of noise to everything such as the shape of the tree structure, allocated processing power. When the analysis is finished it will then compare the score results with other versions to learn which weighting system works the best for the given situation.

Does that answer your questions?

Well, your answer gave me interesting insights into your AI. Thanks a lot. I still have some questions if you don't mind.

Is your weighting based optimization algorithm related to any popular optimization method (e.g. evolutionary algorithms), e.i. did you take inspirations from popular optimization methods?

Under which circumstances is your AI using a hard-coded algorithm for e.g. processing power allocation and how differs such a hard-coded algorithm from an optimized algorithm w.r.t. the representation? Are you making use of parameter collections to represent optimizable algorithms?

I should add that the minor noise that I'm referring to is close to a gaussian distribution.

Evolutionary strategies are usually also using Gaussian distributed noise for optimization. It is used in the mutation phase with the goal to create new individuals (e.g. algorithms) that could perform better than previous ones.

8

Robotics News / MOVED: Intel debuts Compute Card

« Last post by Freddy on April 25, 2018, 10:47:42 pm »

This topic has been moved to General Hardware Talk.

http://aidreams.co.uk/forum/index.php?topic=13054.0

9

General AI Discussion / Re: Eva Progress, from AGI to ASI

« Last post by unreality on April 25, 2018, 10:18:53 pm »

There are reasons that LOCKSUIT behaves the way he does which you would lack the empathy to understand. As irritating as he can be, we put up with him because he makes an effort, he shows signs of improvement, and he is entertaining. For all that, he is still more mature than you, and he has considerably more credibility. He isn't a liar like you, and he isn't an assh*le like you. If he stopped posting we would miss him. Nobody would miss you.

Still pissed because I pointed out your science error? :/  Anyhow, I'm still waiting for your proof that I'm a liar.

By the way, I'm pretty sure he's not allowed to post someone's private messages without their permission. Looks like both of you have no morals. Posting PM without permission, and calling people a liar based on no evidence. Two of a kind.

10

General AI Discussion / Re: Eva Progress, from AGI to ASI

« Last post by infurl on April 25, 2018, 10:15:09 pm »

There are reasons that LOCKSUIT behaves the way he does which you would lack the empathy to understand. As irritating as he can be, we put up with him because he makes an effort, he shows signs of improvement, and he is entertaining. For all that, he is still more mature than you, and he has considerably more credibility. He isn't a liar like you, and he isn't an assh*le like you. If he stopped posting we would miss him. Nobody would miss you.