![[dr. bob]](/dr-bob-sm.gif)
Dr. Bob is Robert Hsiung, MD,
bob@dr-bob.orgShown: posts 1 to 12 of 12. This is the beginning of the thread.
Posted by alexandra_k on June 25, 2005, at 17:20:40
Perhaps it is impossible to attempt to come up with necessary and sufficient conditions for civility. Even if this is so, it doesn’t rule out the possibility of an automated civility checker. There has been success with training PDP / connectionist networks to perform functions that resist the sort of neat delineation that is required to write a traditional program. Connectionist networks can be simulated on an ordinary computer. They consist of a series of nodes. The nodes are arranged in layers (to constitute layers of processing). Each node has a threshold, and when that threshold is reached the node sends weighted connections to nodes in a subsequent layer.
http://www.mind.ilstu.edu/images/gnnv3layer.jpg
(Of course you need many more nodes than that, and more than one hidden layer. You would also want two output nodes ‘civil’ and ‘uncivil’ for civility determinations).
Training the network consists in feeding in information (posts) and seeing what determination the machine makes. If the determination is correct then a new example is fed in. When the machine makes an incorrect determination there are statistical algorithms for adjusting the node thresholds / weightings of the connections. The net preserves past determinations and adapts so as to produce the correct result on new determinations. The nets performance improves over time. What is interesting about such nets is that they have had a lot more success with areas in which traditional programs stumble. They are much better at being able to generalize to a novel instance of the same type of phenomena. Due to this they are better at tasks such as face recognition, object recognition, etc and one would expect that they would be better able to handle civility determinations.
Automated Asterisking
Currently there is a list of words that are considered uncivil. The automated system just matches words in the post with the words on the list. The automated system can’t recognize novel variants where there is an idiosyncratic mis-spelling (muthaf[*]cka) and it can’t recognize then a swear word is embedded in a greater word (motherf[*]cker). Nets would have better success with generalizing to new cases such as these. If a net was trained in one task (automated asterisking) then it is also capable of having new tasks superimposed over top of the old tasks (so long as you have enough nodes and weighted connections).
Anyway…
That’s what I have been thinking. I was a bit worried that you would need rules for translating posts into the first layer of nodes. I was worried that you would have to decide upon meaningful units (words – sentences – paragraphs) – but office mate says you don’t have to worry about this – you just let the net figure that out itself. Of course… It would be a pretty major project. You would need a pretty grunty computer to deliver determinations in anything approaching real time. It is possible in principle but at the moment it is probably a little like the thought of building a stainless steel ladder to the moon.But…
What I was thinking was that over time the net would probably get pretty good in its performance. What would be interesting is that when people hit to ‘submit’ their posts then at the confirmation step the civility checker could post a message ‘this post has been determined to be uncivil by the civility checker. Please reconsider your post. If you choose not to and Dr Bob considers this post to be uncivil then you choose to accept the consequences of your post’. If the poster posts their post unaltered then the machine could post a copy of its determination over on admin. If you agree with the determination then the poster has been warned already. If you do not agree then that’s ok, we could just ignore the machine (and alter the weightings so it won’t do that again). That way we would have an automated civility buddy too Of course such a process doesn’t help all that much with respect to coming to understand the civility rules / determinations. Even once you have designed a net it remains something of a mystery as to *why* or *how* it works. The net relies on a trainer to determine when the net is correct and when it needs to have its weightings adjusted. If it gets good at handling the easy cases, then we might be more inclined to trust it on the harder ones, however.
Posted by gardenergirl on June 26, 2005, at 0:01:07
In reply to Automated Civility Checker, posted by alexandra_k on June 25, 2005, at 17:20:40
I remember learning about connectionist networks in my Cognition class. Very intriguing and complex. You describe it well.
>
> Of course such a process doesn’t help all that much with respect to coming to understand the civility rules / determinations. Even once you have designed a net it remains something of a mystery as to *why* or *how* it works.Hmmmmm, so the net would just "know it when it sees it?" Tee hee
gg
Posted by so on June 26, 2005, at 0:10:53
In reply to Re: Automated Civility Checker » alexandra_k, posted by gardenergirl on June 26, 2005, at 0:01:07
> > Of course such a process doesn’t help all that much with respect to coming to understand the civility rules / determinations. Even once you have designed a net it remains something of a mystery as to *why* or *how* it works.
>
> Hmmmmm, so the net would just "know it when it sees it?" Tee hee
>
> gg
>
No, such a net would have to be programmed with exact algorithms. Creating the algorithms would require anticipation of use cases, most likely based on prior experience and on speculation. It would only be a mystery to those who don't understand the algorithms, or to algorithm designers who are confronted by inconsistencies that arise when their algorithms are applied.
Posted by gabbii on June 26, 2005, at 7:33:23
In reply to Re: Automated Civility Checker, posted by so on June 26, 2005, at 0:10:53
>
> > > Of course such a process doesn’t help all that much with respect to coming to understand the civility rules / determinations. Even once you have designed a net it remains something of a mystery as to *why* or *how* it works.
> >
> No, such a net would have to be programmed with exact algorithms. Creating the algorithms would require anticipation of use cases, most likely based on prior experience and on speculation. It would only be a mystery to those who don't understand the algorithms, or to algorithm designers who are confronted by inconsistencies that arise when their algorithms are applied.I think it would really annoy me.
I mean, it couldn't tell if you were referring to an outside situation, or to something someone said in your personal life, and not directing the comment toward a poster could it?For example, saying "it's causing a stir because it's so offensive" (referring to say, a situation in the news) would probably be considered "uncivil" wouldn't it?
Posted by alexandra_k on June 26, 2005, at 8:00:19
In reply to Re: Automated Civility Checker, posted by so on June 26, 2005, at 0:10:53
I don't know too much about them so this might be misinformation but...
I don't think that it is that they are programmed with exact algorithms. I think the algorithms are used to adjust the weightings / thresholds when the determination is incorrect. There are algorithms available. I'm not sure how specific the algorithms are to architecture (with respect to number of nodes and number of layers and whether there are one or two or more output nodes). I'm not sure if it is relevant at all what sort of task the net is performing (object recognition vs civility determination for instance).
Along the lines of building a stainless steel ladder to the moon... I was thinking it would be kind of interesting if there was a Dr Bob checker (where he determined whether the weightings needed to be adjusted or not) and if there was another poster checker where posters votes could determine whether the weightings get adjusted or not.
Then you could compare them... Compare the weightings and the thresholds... And worry about how to 'translate' the math and worry about whether there might just be some kind of rule lurking implicitly in there after all...
Most probably...
Such machines lack semantics. They lack meanings. They lack understanding. Searle "Rediscovery of the Mind" thinks they lack it as a matter of principle. AI will never have intentional (mental) content (meaningful thoughts about the world).
Other theorists disagree. Some say that if you hooked a net up to appropriate inputs (sensory information such as sight, hearing, etc) and appropriate outputs (some kind of body to move around and interact with the world) then you will get mental content after all.
I don't know how good performance could get...
How many nodes (neurons) are there in the brain???
Lots.
But we don't need all of them...
So you could think of it as a relatively coarse grained simulation of one functional mechanism in the mind / brain.
If turing machine functionalism is correct and consciousness just is the experience of certain kinds of complex causal interaction then it is even possible that such a machine could have conscious states.
Imagine that. An eternity of civility determinations on our posts
NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
Surely there must be other ways to achieve immortality :-)I don't know how good performance can get.
I think people are playing around with different architectures. I'd have a go at drawing some but don't really know how to attach pictures...
Posted by Dinah on June 26, 2005, at 9:16:08
In reply to Automated Civility Checker, posted by alexandra_k on June 25, 2005, at 17:20:40
My response would be a shudder, and recoil.
I get all the amusement I need from the automated curse checker.
I don't think I would get nearly as much amusement from an automated program that checked word sequences without the leavening of human judgement.
Just my two cents.
Posted by so on June 26, 2005, at 10:54:20
In reply to Re: Automated Civility Checker » so, posted by gabbii on June 26, 2005, at 7:33:23
> I think it would really annoy me.
> I mean, it couldn't tell if you were referring to an outside situation, or to something someone said in your personal life, and not directing the comment toward a poster could it?
>
> For example, saying "it's causing a stir because it's so offensive" (referring to say, a situation in the news) would probably be considered "uncivil" wouldn't it?Algorithms could eventually be designed to detect context - it's mostly a matter of processor time required to check context, especially of nouns and of prepositional phrases. Of course, English does allow considerable ambiguity in the meaning of words and usage. English is a loose language and most average English writers don't write with sufficient precision to satisfy even the grammar checkers now available.
An algorithm could be written to check compliance with terms of service, if they were specific. If it were a tool available for voluntary use, and not the last word, it could be useful for people trying to comply with exceptionally unique terms of service such as those implemented on at least one mental health mutual support site. The problem I foresee arises with one individual defining their local terms of service as "civility." I doubt any machine can correct cognitive disonnance introduced in the minds of a community by a powerful individual's decision to impose their own definition for a word so descriptive of basic human behavior.
Posted by gabbii on June 26, 2005, at 12:15:45
In reply to Re: Automated Civility Checker » gabbii, posted by so on June 26, 2005, at 10:54:20
Posted by gabbii on June 26, 2005, at 12:20:58
In reply to Re: Automated Civility Checker, posted by Dinah on June 26, 2005, at 9:16:08
> My response would be a shudder, and recoil.
>
> I get all the amusement I need from the automated curse checker.
>
> I don't think I would get nearly as much amusement from an automated program that checked word sequences without the leavening of human judgement.
>
> Just my two cents.I agree, if it ever is implemented I hope it's optional. I'm thinking of the extra time and frustration trying to re-word your post, not because they are uncivil, but because you know the civility checker will see it that way (we know there will be many, many mistakes) and you don't want it spat back at you.. well it sounds exceedingly frustrating.
Posted by alexandra_k on June 26, 2005, at 18:08:31
In reply to Re: Automated Civility Checker, posted by Dinah on June 26, 2005, at 9:16:08
It would probably make me shudder if it delivered wrong result after wrong result... If it got quite good then I guess I'd be pretty excited about what it could do :-)
Likewise, Dr Bob would have to decide whether it delivered the correct or incorrect result (and thus the weightings / thresholds would need to be adjusted).
If you thought it wasn't correct when you went to confirm your post you could ignore it, of course. Like how the current system unnecessarily askerisks p*m p*ms.
You would only want to use it as a civility buddy if it didn't do too many false alarms. You would want it to be performing well before inflicting it on posters in that way.
But asterisking...
Asterisking could be a good place to start.
Posted by Dr. Bob on June 27, 2005, at 0:00:47
In reply to Automated Civility Checker, posted by alexandra_k on June 25, 2005, at 17:20:40
[ redirected from: http://www.dr-bob.org/babble/admin/20050614/msgs/513223.html ]
> What is interesting about such nets is that they have had a lot more success with areas in which traditional programs stumble.
>
> Of course… It would be a pretty major project.I agree, it would be really interesting. But it's way beyond me!
Bob
Posted by alexandra_k on June 27, 2005, at 2:09:48
In reply to Re: Automated Civility Checker, posted by Dr. Bob on June 27, 2005, at 0:00:47
deja vu.
Sorry bout that.
This is the end of the thread.
Psycho-Babble Administration | Extras | FAQ
Dr. Bob is Robert Hsiung, MD,
bob@dr-bob.org
Script revised: February 4, 2008
URL: http://www.dr-bob.org/cgi-bin/pb/mget.pl
Copyright 2006-17 Robert Hsiung.
Owned and operated by Dr. Bob LLC and not the University of Chicago.