Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > comp.theory > #25729 > unrolled thread
| Started by | "dklei...@gmail.com" <dkleinecke@gmail.com> |
|---|---|
| First post | 2020-11-23 11:37 -0800 |
| Last post | 2020-11-24 03:59 +0000 |
| Articles | 5 on this page of 45 — 7 participants |
Back to article view | Back to comp.theory
Not about Olcutt "dklei...@gmail.com" <dkleinecke@gmail.com> - 2020-11-23 11:37 -0800
Re: Not about Olcutt olcott <NoOne@NoWhere.com> - 2020-11-23 13:46 -0600
Re: Not about Olcutt Richard Damon <Richard@Damon-Family.org> - 2020-11-23 15:04 -0500
Re: Not about Olcutt Jeff Barnett <jbb@notatt.com> - 2020-11-23 15:52 -0700
Re: Not about Olcutt Richard Damon <Richard@Damon-Family.org> - 2020-11-23 18:08 -0500
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-23 17:35 -0600
Re: Not about Olcutt Andy Walker <anw@cuboid.co.uk> - 2020-11-24 10:36 +0000
Re: Not about Olcutt Malcolm McLean <malcolm.arthur.mclean@gmail.com> - 2020-11-23 13:06 -0800
Re: Not about Olcutt "dklei...@gmail.com" <dkleinecke@gmail.com> - 2020-11-23 13:51 -0800
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-23 15:55 -0600
Re: Church-Turing thesis Jeff Barnett <jbb@notatt.com> - 2020-11-23 16:16 -0700
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-23 17:34 -0600
Re: Church-Turing thesis Richard Damon <Richard@Damon-Family.org> - 2020-11-23 19:51 -0500
Re: Church-Turing thesis Jeff Barnett <jbb@notatt.com> - 2020-11-23 18:10 -0700
Re: Church-Turing thesis Jeff Barnett <jbb@notatt.com> - 2020-11-24 14:11 -0700
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-24 15:40 -0600
Re: Church-Turing thesis Malcolm McLean <malcolm.arthur.mclean@gmail.com> - 2020-11-24 15:12 -0800
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-24 17:28 -0600
Re: Church-Turing thesis Richard Damon <Richard@Damon-Family.org> - 2020-11-24 18:43 -0500
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-24 18:05 -0600
Re: Church-Turing thesis Ben Bacarisse <ben.usenet@bsb.me.uk> - 2020-11-25 03:30 +0000
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-24 21:43 -0600
Re: Church-Turing thesis Ben Bacarisse <ben.usenet@bsb.me.uk> - 2020-11-25 21:25 +0000
Re: Church-Turing thesis Jeff Barnett <jbb@notatt.com> - 2020-11-24 16:26 -0700
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-24 17:33 -0600
Re: Church-Turing thesis Richard Damon <Richard@Damon-Family.org> - 2020-11-24 18:38 -0500
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-24 17:44 -0600
Re: Church-Turing thesis Jeff Barnett <jbb@notatt.com> - 2020-11-24 16:19 -0700
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-24 17:26 -0600
Re: Church-Turing thesis Jeff Barnett <jbb@notatt.com> - 2020-11-24 18:12 -0700
Re: Church-Turing thesis Ben Bacarisse <ben.usenet@bsb.me.uk> - 2020-11-25 03:33 +0000
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-24 21:47 -0600
Re: Church-Turing thesis Ben Bacarisse <ben.usenet@bsb.me.uk> - 2020-11-24 23:23 +0000
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-24 17:31 -0600
Re: Church-Turing thesis Jeff Barnett <jbb@notatt.com> - 2020-11-24 18:14 -0700
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-24 19:37 -0600
Re: Church-Turing thesis Jeff Barnett <jbb@notatt.com> - 2020-11-25 11:49 -0700
Re: Church-Turing thesis Ben Bacarisse <ben.usenet@bsb.me.uk> - 2020-11-25 21:30 +0000
Re: Not about Olcutt Richard Damon <Richard@Damon-Family.org> - 2020-11-23 17:02 -0500
Re: Not about Olcutt Jeff Barnett <jbb@notatt.com> - 2020-11-23 15:56 -0700
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-23 17:36 -0600
Re: Not about Olcutt Jeff Barnett <jbb@notatt.com> - 2020-11-23 15:54 -0700
Re: Not about Olcutt Ben Bacarisse <ben.usenet@bsb.me.uk> - 2020-11-24 02:00 +0000
Re: Church-Turing thesis olcott <NoOne@NoWhere.com> - 2020-11-23 20:12 -0600
Re: Church-Turing thesis Ben Bacarisse <ben.usenet@bsb.me.uk> - 2020-11-24 03:59 +0000
Page 3 of 3 — ← Prev page 1 2 [3]
| From | olcott <NoOne@NoWhere.com> |
|---|---|
| Date | 2020-11-23 17:36 -0600 |
| Subject | Re: Church-Turing thesis |
| Message-ID | <RoKdnaQVNpAJ2iHCnZ2dnUU7-W-dnZ2d@giganews.com> |
| In reply to | #25752 |
On 11/23/2020 4:56 PM, Jeff Barnett wrote: > On 11/23/2020 2:51 PM, dklei...@gmail.com wrote: >> On Monday, November 23, 2020 at 1:06:29 PM UTC-8, >> malcolm.ar...@gmail.com wrote: >>> On Monday, 23 November 2020 at 19:37:13 UTC, dklei...@gmail.com wrote: >>>> I hope that comp.theory is the best place for the following >>>> question: >>>> >>>> Has it ever been shown that a Turing Machine can do the >>>> kind of neural network calculating done by AI such as Siri? >>>> >>>> This in the context of whether or not such calculating can >>>> legitimately be called algorithmic. It is not algorithmic in >>>> sense Knuth uses the word "algorithm". >>>> >>> "Neural network" was a great marketing term. But all a neural >>> network is is a directed graph, weights, and a sigmoid function to >>> prevent high weights pushing output very high and to make the >>> results continously differentiable to help out the training algorithm. >>> >>> It doesn't emulate a real human brain. >> >> The problem here is that the neural networks continually modify the >> network and a Turing Machine can't be modified on the fly. Can we >> use a Turing Machine to simulate another mode of calculating >> (neural nets) and still say that Turing Machine is doing the processing. >> That is, assuming a Turing Machine can simulate a neural net. The >> Church-Turing thesis is not a theorem - how sure of it are we? > > A C neural net program keeps its net in storage; a TM neural net keeps > its net on its tape. So what's the problem with modifying the tape? > None! That's mostly what TMs do. Good point. That did not intitally occur to me. -- Copyright 2020 Pete Olcott "Great spirits have always encountered violent opposition from mediocre minds." Einstein
[toc] | [prev] | [next] | [standalone]
| From | Jeff Barnett <jbb@notatt.com> |
|---|---|
| Date | 2020-11-23 15:54 -0700 |
| Message-ID | <rpheir$ojf$2@dont-email.me> |
| In reply to | #25736 |
On 11/23/2020 2:06 PM, Malcolm McLean wrote: > On Monday, 23 November 2020 at 19:37:13 UTC, dklei...@gmail.com wrote: >> I hope that comp.theory is the best place for the following >> question: >> >> Has it ever been shown that a Turing Machine can do the >> kind of neural network calculating done by AI such as Siri? >> >> This in the context of whether or not such calculating can >> legitimately be called algorithmic. It is not algorithmic in >> sense Knuth uses the word "algorithm". >> > "Neural network" was a great marketing term. But all a neural > network is is a directed graph, weights, and a sigmoid function to > prevent high weights pushing output very high and to make the > results continously differentiable to help out the training algorithm. > > It doesn't emulate a real human brain. Even if it did, a TM could implement one as easily as a C system. -- Jeff Barnett
[toc] | [prev] | [next] | [standalone]
| From | Ben Bacarisse <ben.usenet@bsb.me.uk> |
|---|---|
| Date | 2020-11-24 02:00 +0000 |
| Message-ID | <875z5vqzs4.fsf@bsb.me.uk> |
| In reply to | #25729 |
"dklei...@gmail.com" <dkleinecke@gmail.com> writes: > I hope that comp.theory is the best place for the following > question: > > Has it ever been shown that a Turing Machine can do the > kind of neural network calculating done by AI such as Siri? > > This in the context of whether or not such calculating can > legitimately be called algorithmic. It is not algorithmic in > sense Knuth uses the word "algorithm". The only contentious elements would be the use of random data. If your neural net can use pseudo-random numbers (and many NN simulators do just that of course) then there is nothing "un-algorithmic" about a neural net. Then again, there are TM models that include access to random data (usually on a second tape), so such a model could be used as the theoretical basis for a neural net with a physically random data source. -- Ben.
[toc] | [prev] | [next] | [standalone]
| From | olcott <NoOne@NoWhere.com> |
|---|---|
| Date | 2020-11-23 20:12 -0600 |
| Subject | Re: Church-Turing thesis |
| Message-ID | <c9-dnWvSZPFv9iHCnZ2dnUU7-QvNnZ2d@giganews.com> |
| In reply to | #25770 |
On 11/23/2020 8:00 PM, Ben Bacarisse wrote: > "dklei...@gmail.com" <dkleinecke@gmail.com> writes: > >> I hope that comp.theory is the best place for the following >> question: >> >> Has it ever been shown that a Turing Machine can do the >> kind of neural network calculating done by AI such as Siri? >> >> This in the context of whether or not such calculating can >> legitimately be called algorithmic. It is not algorithmic in >> sense Knuth uses the word "algorithm". > > The only contentious elements would be the use of random data. If your > neural net can use pseudo-random numbers (and many NN simulators do just > that of course) then there is nothing "un-algorithmic" about a neural > net. > > Then again, there are TM models that include access to random data > (usually on a second tape), so such a model could be used as the > theoretical basis for a neural net with a physically random data > source. > Yes hardware random number generators are not any aspect of a Turing machine. https://en.wikipedia.org/wiki/Hardware_random_number_generator -- Copyright 2020 Pete Olcott "Great spirits have always encountered violent opposition from mediocre minds." Einstein
[toc] | [prev] | [next] | [standalone]
| From | Ben Bacarisse <ben.usenet@bsb.me.uk> |
|---|---|
| Date | 2020-11-24 03:59 +0000 |
| Subject | Re: Church-Turing thesis |
| Message-ID | <87d003pfqe.fsf@bsb.me.uk> |
| In reply to | #25771 |
olcott <NoOne@NoWhere.com> writes: > On 11/23/2020 8:00 PM, Ben Bacarisse wrote: >> "dklei...@gmail.com" <dkleinecke@gmail.com> writes: >> >>> I hope that comp.theory is the best place for the following >>> question: >>> >>> Has it ever been shown that a Turing Machine can do the >>> kind of neural network calculating done by AI such as Siri? >>> >>> This in the context of whether or not such calculating can >>> legitimately be called algorithmic. It is not algorithmic in >>> sense Knuth uses the word "algorithm". >> >> The only contentious elements would be the use of random data. If your >> neural net can use pseudo-random numbers (and many NN simulators do just >> that of course) then there is nothing "un-algorithmic" about a neural >> net. >> >> Then again, there are TM models that include access to random data >> (usually on a second tape), so such a model could be used as the >> theoretical basis for a neural net with a physically random data >> source. > > Yes hardware random number generators are not any aspect of a Turing > machine. Hardware is not any aspect of a Turing machine. Genuinely random numbers can be if you choose to include them in the mathematical specification. -- Ben.
[toc] | [prev] | [standalone]
Page 3 of 3 — ← Prev page 1 2 [3]
Back to top | Article view | comp.theory
csiph-web