On YouTube: Genifer in 4 minutes

[YKY (Yan King Yin, 甄景贤)]

The purpose of this is just to give a general overview of Genifer:

http://www.youtube.com/watch?v=SurJZXEX3B8

Any request for videos on specific topics? =)

-- 

KY

"The ultimate goal of mathematics is to eliminate any need for intelligent thought" -- Alfred North Whitehead

[swkane]

Yes,

   Let's see a video on the specifics of the inductive learner. I'm open minded but still quite skeptical on how effective Genifer's inductive learner will be for real world problems.

Steven

2011/12/9 YKY (Yan King Yin, 甄景贤) <generic.in...@gmail.com>

[YKY (Yan King Yin, 甄景贤)]

On Sun, Dec 11, 2011 at 3:20 AM, swkane <diss...@gmail.com> wrote:

Yes,

   Let's see a video on the specifics of the inductive learner. I'm open minded but still quite skeptical on how effective Genifer's inductive learner will be for real world problems.

Steven

I already have a whole bunch of ideas of optimizing inductive learning in the slides (and some in my head).  I was on the way to coming up with even better ideas.  But, the current situation is that OpenCog is a functioning AGI and is grabbing all the low-hanging fruits, and draining my potential revenues and sources of funding (Ben is getting funded in Hong Kong and China via the very same agencies that could have funded me).  It seems that if I come up with a cool idea for inductive learning, Opencog would copy it also -- they have done that twice already (1. fuzzy-probabilistic logic, 2. hierarchical clustering for inference optimization).

What should I do?

KY

[Matt Mahoney]

2011/12/10 YKY (Yan King Yin, 甄景贤) <generic.in...@gmail.com>:

You can steal their ideas. OpenCog is open source. Download the
software and improve it.

BTW, YouTube says the video was removed by the owner so I didn't get to see it.

--
-- Matt Mahoney, mattma...@gmail.com

[YKY (Yan King Yin, 甄景贤)]

On Sun, Dec 11, 2011 at 8:57 AM, Matt Mahoney <mattma...@gmail.com> wrote:

You can steal their ideas. OpenCog is open source. Download the
software and improve it.

The problem is that they don't have any ideas worth stealing -- we have similar logical foundations, so perhaps I have already internalized Ben's ideas from many years ago.  One of their special idea is Atomspace, but I'm still unsure if I want to use it.

The current situation is similar to Windows copying Apple's GUI idea in the 80s =(

BTW, YouTube says the video was removed by the owner so I didn't get to see it.

Sorry, I revised it, so the URL changed: 

Genifer in 4 minutes

And this is just for fun:

The 2nd law of thermodynamics

KY

[Matt Mahoney]

2011/12/10 YKY (Yan King Yin, 甄景贤) <generic.in...@gmail.com>:

> On Sun, Dec 11, 2011 at 8:57 AM, Matt Mahoney <mattma...@gmail.com>
> wrote:
>>
>> You can steal their ideas. OpenCog is open source. Download the
>> software and improve it.
>
> The problem is that they don't have any ideas worth stealing -- we have
> similar logical foundations, so perhaps I have already internalized Ben's
> ideas from many years ago.  One of their special idea is Atomspace, but I'm
> still unsure if I want to use it.

They probably feel the same way about your ideas, so I wouldn't worry about it.

> The current situation is similar to Windows copying Apple's GUI idea in the
> 80s =(

Microsoft only did that after Apple proved it would work. (And Apple
stole the ideas from Xerox). Nobody is going to steal any AI ideas
because Genifer and OpenCog are both a long long way from doing
anything useful. The best you can do now is publish your ideas and
hope that someone steals them so you will at least have a reputation
for coming up with good ideas.

>> BTW, YouTube says the video was removed by the owner so I didn't get to
>> see it.
>
> Sorry, I revised it, so the URL changed:
> Genifer in 4 minutes

I think that's a nice summary. But it is not obvious to me how to
build the KB other than write the rules by hand. To automate the
process, Genifer would need to understand natural language. If it
could do that, then we would have already solved the AI problem.

> And this is just for fun:
> The 2nd law of thermodynamics

Yes, the missing ingredient is that uncertainty (what entropy
measures) is only defined for observers. An observer is any agent that
performs time-irreversible operations such as writing to memory.
Recording instruments, computers, and human brains are observers.
Observers don't exist in the complete (time-reversible wave model)
model of physics that describes our universe. Observers only exist in
the incomplete models of physics that exists in the minds of
observers.

So yes, it is possible for Maxwell's Demon to exist. But it is not
possible to observe it. Any process described as moving backward in
time would be an example.

-- Matt Mahoney, mattma...@gmail.com

[YKY (Yan King Yin, 甄景贤)]

By the way, someone complained about the female computer voice, saying that it's annoying.  I used it because I tried narrating it myself but it came out badly (I need more practice).  I'm now finding an English-speaking female student to narrate it, but she's having final exams now.  Also, I will eliminate the first-person perspective to avoid confusion.

This brings up the issue of Genifer's "gender".  I want to clarify that Genifer is intended as an intelligent tool but we do not intend it to have sentience (emotions, consciousness, self-identity, etc).  Therefore the issue of gender is moot.  Also, I don't think it should have an a priori gender bias in its knowledge base, other than what the user inputs.  Genifer is just a code name for this project that I find to be cute / clever.  In the next version I'd de-emphasize this name.

KY

[Matt Mahoney]

I dunno. I kind of liked the synthesized voice. I mean, it's
appropriate for an AI, isn't it?

2011/12/11 YKY (Yan King Yin, 甄景贤) <generic.in...@gmail.com>:

--
-- Matt Mahoney, mattma...@gmail.com