On Thu, May 24, 2018 at 4:04 AM, Alexey Potapov wrote:
> Linas,
> OK, let's continue the discussion about probabilities.
>
> 2018-05-22 21:18 GMT+03:00 Linas Vepstas :
>
>> Do not assume that a probability is what you actually want. Let me give
>> three examples.
>>
>> In real life, when you see
Linas,
OK, let's continue the discussion about probabilities.
2018-05-22 21:18 GMT+03:00 Linas Vepstas :
> Do not assume that a probability is what you actually want. Let me give
> three examples.
>
> In real life, when you see a crow, and it is dark, and you want to talk
> about it, you just sa
On Wed, May 23, 2018 at 1:01 PM, Alexey Potapov wrote:
> 2018-05-22 21:18 GMT+03:00 Linas Vepstas :
>
>>
>>
>> On Sun, May 20, 2018 at 3:26 AM, Alexey Potapov
>> wrote:
>>
>>>
>>>
>>> For me, observational data is sensory data. It doesn't contain concepts,
>>> predicates, etc. . If we have an ob
2018-05-22 21:18 GMT+03:00 Linas Vepstas :
>
>
> On Sun, May 20, 2018 at 3:26 AM, Alexey Potapov
> wrote:
>
>>
>>
>> For me, observational data is sensory data. It doesn't contain concepts,
>> predicates, etc. . If we have an observation that a particular crow is
>> black, ... But there are no pu
2018-05-22 1:54 GMT+03:00 Linas Vepstas :
> On Sat, May 19, 2018 at 1:00 PM, Alexey Potapov
> wrote:
>
>> Well... traditional probabilistic programming is a logical probabilistic
>> programming. It's definitely not about lambda-calculus.
>>
>
> I don't know what to do with this statement. There i
On Wed, May 16, 2018 at 6:32 PM, Ben Goertzel wrote:
> Alexey, Nil, Zar, Linas, others...
>
>
> GENERAL BLATHER
>
> 2) A route with a large role for probabilistic-logic theorem-proving
> is one viable route
>
I'm starting to wonder if probabilistic logic, in this narrow sense, is
actually needed
On Sun, May 20, 2018 at 3:26 AM, Alexey Potapov wrote:
>
>
> For me, observational data is sensory data. It doesn't contain concepts,
> predicates, etc. . If we have an observation that a particular crow is
> black, ... But there are no purely black crows. It's just an abstraction,
> which itself
2018-05-22 7:04 GMT+03:00 Nil Geisweiller :
>
> Sounds interesting. Without details being provided that sounds to me like
> meta-learning + schematization (turning the most frequent paths of a
> generalized solver into a more narrow efficient program, as you did mention
> during the last Singularit
On 05/21/2018 11:19 PM, Alexey Potapov wrote:
This might be ok when we are talking about small-dimensional tasks, but
I don't think this is a good idea for real-world problems...
Yeah, could be. Or I suppose it could be some hybrid incremental/batch.
BTW, one my colleague (Vitaly Khudobahshov)
On Sat, May 19, 2018 at 1:00 PM, Alexey Potapov wrote:
>
>
> Well... traditional probabilistic programming is a logical probabilistic
> programming. It's definitely not about lambda-calculus.
>
I don't know what to do with this statement. There is a famous theorem, the
church-turing theorem, dat
Nil,
>
> Deduction system can be understood very broadly, and may encompass
> inferences based on PPL models as well.
>
> PLN definitely draws, at least in principle, the relationship between
> deduction and data.
>
> ATM in practice it's a bit lacking though, for instance the link between
> the
Yeah, true...
On Mon, 21 May 2018, 13:41 Nil Geisweiller, wrote:
> On 05/21/2018 01:48 PM, Ben Goertzel wrote:
> > But Nil -- those record-keeping links can be put in an auxiliary
> > Atomspace, not necessarily the same Atomspace where the main thrust of
> > reasoning is proceeding...
>
> Yes, b
On 05/21/2018 01:48 PM, Ben Goertzel wrote:
But Nil -- those record-keeping links can be put in an auxiliary
Atomspace, not necessarily the same Atomspace where the main thrust of
reasoning is proceeding...
Yes, but for rules like incremental direct calculation, and TV revision
in general, it
But Nil -- those record-keeping links can be put in an auxiliary
Atomspace, not necessarily the same Atomspace where the main thrust of
reasoning is proceeding...
On Mon, May 21, 2018 at 8:45 AM, 'Nil Geisweiller' via opencog
wrote:
> On 05/19/2018 09:00 PM, Alexey Potapov wrote:
>>
>> Our knowle
On 05/19/2018 09:00 PM, Alexey Potapov wrote:
Our knowledge is built from data. Deduction systems (probabilistic or
not) lack this connection, while functional PPLs are well-suited for this.
Deduction system can be understood very broadly, and may encompass
inferences based on PPL models as we
Alexey -- e.g. if one stays in the world of finite discrete
distributions, one can construct probabilistic logics with
sampling-based semantics...
https://arxiv.org/pdf/1602.06420.pdf
To extend this to deal with PLN, basically one just needs to jump up
to second (and for quantifiers, third) order
> But how will you calculate P(image|crow,black)?
Well as you know, if you really want to, something like "the RGB value
of the pixel at coordinate (444,555) is within a distance .01 of
(.3,.7,.8)" can be represented as a logical atom ... so there is no
problem using logic to reason about percept
Ben,
2018-05-20 8:04 GMT+03:00 Ben Goertzel :
> Alexey,
>
> ***
> Our knowledge is built from data. Deduction systems (probabilistic or
> not) lack this connection, while functional PPLs are well-suited for
> this.
> ***
>
> I don't understand why you think this way...
>
> The semantics of prob
Alexey,
***
Our knowledge is built from data. Deduction systems (probabilistic or
not) lack this connection, while functional PPLs are well-suited for
this.
***
I don't understand why you think this way...
The semantics of probabilistic logic systems can be naturally framed
in a fully observatio
>
> The difference between a theorem proving
> based AI and a program learning based AI is merely an “implementation
> detail” ;-) …
>
Well, true, but the devil is in the implementation detail.
--
You received this message because you are subscribed to the Google Groups
"opencog" group.
To unsu
>
> I am very much trying to pursue a probabilistic approach. The question is
> then "probabilities on what?"
>
Ultimately, probabilities over observational data.
>
> You know probabilistic programming very very much better than I, so please
> correct my mis-steps. In probabilistic programming,
Alexey, Nil, Zar, Linas, others...
Ah, I have some new thoughts on the "theorem proving + AGI + parsing"
side, but will have to wait to type them in till I get a little time
at the computer, I'm traveling between meetings and conferences in
Europe just now...
And here we go...
GENERAL
Ah, I have some new thoughts on the "theorem proving + AGI + parsing"
side, but will have to wait to type them in till I get a little time
at the computer, I'm traveling between meetings and conferences in
Europe just now...
On Tue, May 15, 2018 at 12:44 AM, Linas Vepstas wrote:
>
>
> On Mon, M
On Mon, May 14, 2018 at 4:54 PM, Alexey Potapov wrote:
> Hi Linas,
> this is quite an interesting discussion, and I believe we should involve
> Ben and others in it.
>
Ben has been involved in this discussion for a decade; I think he knows
the general outline, even as we argue violently about t
24 matches
Mail list logo
