Re: [agi] new paper: What Do You Mean by AI?
Pei, you classified NARS as a principle-based AI. Are there any others in that category? What about Novamente? -- Matt Mahoney, [EMAIL PROTECTED] - Original Message From: Pei Wang [EMAIL PROTECTED] To: agi@v2.listbox.com agi@v2.listbox.com Sent: Friday, November 17, 2006 11:51:58 AM Subject: [agi] new paper: What Do You Mean by AI? Hi, A new paper of mine is put on-line for comment. English corrections are also welcome. You can either post to this mailing list or send me private emails. Thanks in advance. Pei --- TITLE: What Do You Mean by AI? ABSTRACT: Many problems in AI study can be traced back to the confusion of different research goals. In this paper, five typical ways to define AI are clarified, analyzed, and compared. It is argued that though they are all legitimate research goals, they lead the research to very different directions. Furthermore, most of them have trouble to give AI a proper identity. URL: http://nars.wang.googlepages.com/wang.AI_Definitions.pdf - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] new paper: What Do You Mean by AI?
Of course, I know many people won't agree with my classification, and many projects are pursuing multiple goals, according to this classification. One example is Novamente --- Ben wants (almost) everything. ;-) In the following webpage I just updated, you can see where I put some representative AGI-related projects in this classification, according to their main goals, as far as I can tell: http://www.cis.temple.edu/~pwang/203-AI/Lecture/AGI.htm Pei On 11/18/06, Matt Mahoney [EMAIL PROTECTED] wrote: Pei, you classified NARS as a principle-based AI. Are there any others in that category? What about Novamente? -- Matt Mahoney, [EMAIL PROTECTED] - Original Message From: Pei Wang [EMAIL PROTECTED] To: agi@v2.listbox.com agi@v2.listbox.com Sent: Friday, November 17, 2006 11:51:58 AM Subject: [agi] new paper: What Do You Mean by AI? Hi, A new paper of mine is put on-line for comment. English corrections are also welcome. You can either post to this mailing list or send me private emails. Thanks in advance. Pei --- TITLE: What Do You Mean by AI? ABSTRACT: Many problems in AI study can be traced back to the confusion of different research goals. In this paper, five typical ways to define AI are clarified, analyzed, and compared. It is argued that though they are all legitimate research goals, they lead the research to very different directions. Furthermore, most of them have trouble to give AI a proper identity. URL: http://nars.wang.googlepages.com/wang.AI_Definitions.pdf - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: Re: [agi] new paper: What Do You Mean by AI?
According to your classification, structure (e.g., to build brain models) behavior (e.g., to simulate human mind) capability (e.g., to solve hard problems) function (e.g., to have cognitive facilities) principle (e.g., to be adaptive and rational) Novamente is based on the final 3 categories, but not the first two. Novamente is not intended to model the human brain or mind, though it draws inspiration from both here and there, as seems seful. In your article you associate Novamente with capability and function but not principle; however, strict adherence to probabilistic semantics is an important aspect of Novamente, and one thing that distinguishes it from its predecessor, Webmind. I believe this is an example of what you call principle. -- Ben On 11/18/06, Pei Wang [EMAIL PROTECTED] wrote: Of course, I know many people won't agree with my classification, and many projects are pursuing multiple goals, according to this classification. One example is Novamente --- Ben wants (almost) everything. ;-) In the following webpage I just updated, you can see where I put some representative AGI-related projects in this classification, according to their main goals, as far as I can tell: http://www.cis.temple.edu/~pwang/203-AI/Lecture/AGI.htm Pei On 11/18/06, Matt Mahoney [EMAIL PROTECTED] wrote: Pei, you classified NARS as a principle-based AI. Are there any others in that category? What about Novamente? -- Matt Mahoney, [EMAIL PROTECTED] - Original Message From: Pei Wang [EMAIL PROTECTED] To: agi@v2.listbox.com agi@v2.listbox.com Sent: Friday, November 17, 2006 11:51:58 AM Subject: [agi] new paper: What Do You Mean by AI? Hi, A new paper of mine is put on-line for comment. English corrections are also welcome. You can either post to this mailing list or send me private emails. Thanks in advance. Pei --- TITLE: What Do You Mean by AI? ABSTRACT: Many problems in AI study can be traced back to the confusion of different research goals. In this paper, five typical ways to define AI are clarified, analyzed, and compared. It is argued that though they are all legitimate research goals, they lead the research to very different directions. Furthermore, most of them have trouble to give AI a proper identity. URL: http://nars.wang.googlepages.com/wang.AI_Definitions.pdf - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: Re: [agi] new paper: What Do You Mean by AI?
On 11/18/06, Ben Goertzel [EMAIL PROTECTED] wrote: According to your classification, structure (e.g., to build brain models) behavior (e.g., to simulate human mind) capability (e.g., to solve hard problems) function (e.g., to have cognitive facilities) principle (e.g., to be adaptive and rational) Novamente is based on the final 3 categories, but not the first two. Novamente is not intended to model the human brain or mind, though it draws inspiration from both here and there, as seems seful. Agree. In your article you associate Novamente with capability and function but not principle; however, strict adherence to probabilistic semantics is an important aspect of Novamente, and one thing that distinguishes it from its predecessor, Webmind. I believe this is an example of what you call principle. Agree, too --- that is why I said you want almost everything. However, whenever a design decision is made, you usually consider more about the system's problem-solving ability, and less about the consistency of its theoretical foundation --- of course, you may argue that it don't conflict with your principle, though I'm not that sure. On the contrary, I'm more sensitive to the theoretical consistency of a system than to its practical ability. Pei -- Ben On 11/18/06, Pei Wang [EMAIL PROTECTED] wrote: Of course, I know many people won't agree with my classification, and many projects are pursuing multiple goals, according to this classification. One example is Novamente --- Ben wants (almost) everything. ;-) In the following webpage I just updated, you can see where I put some representative AGI-related projects in this classification, according to their main goals, as far as I can tell: http://www.cis.temple.edu/~pwang/203-AI/Lecture/AGI.htm Pei On 11/18/06, Matt Mahoney [EMAIL PROTECTED] wrote: Pei, you classified NARS as a principle-based AI. Are there any others in that category? What about Novamente? -- Matt Mahoney, [EMAIL PROTECTED] - Original Message From: Pei Wang [EMAIL PROTECTED] To: agi@v2.listbox.com agi@v2.listbox.com Sent: Friday, November 17, 2006 11:51:58 AM Subject: [agi] new paper: What Do You Mean by AI? Hi, A new paper of mine is put on-line for comment. English corrections are also welcome. You can either post to this mailing list or send me private emails. Thanks in advance. Pei --- TITLE: What Do You Mean by AI? ABSTRACT: Many problems in AI study can be traced back to the confusion of different research goals. In this paper, five typical ways to define AI are clarified, analyzed, and compared. It is argued that though they are all legitimate research goals, they lead the research to very different directions. Furthermore, most of them have trouble to give AI a proper identity. URL: http://nars.wang.googlepages.com/wang.AI_Definitions.pdf - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: Re: Re: [agi] new paper: What Do You Mean by AI?
Agree, too --- that is why I said you want almost everything. However, whenever a design decision is made, you usually consider more about the system's problem-solving ability, and less about the consistency of its theoretical foundation --- of course, you may argue that it don't conflict with your principle, though I'm not that sure. On the contrary, I'm more sensitive to the theoretical consistency of a system than to its practical ability. Pei Hi, Pei ... I think your comments above are more apropos to Webmind than to Novamente. For instance, in my bioinformatics and NLP narrow-AI work I have been extremely impressed with SVM's for supervised categorization. So, in the Webmind philosophy, we would have integrated SVM's into Webmind to serve inside a supervised categorization agent. In Novamente, OTOH, we use MOSES (probabilistic evolutionary learning) for supervised categorization internally (e.g. for finding rules regarding what properties characterize the extension of some concept) ... even though it is not always as accurate as SVM's. The reasons are twofold: a) keeping the overall design as simple as possible b) adherence to the principle of probabilistic semantics (yes, SVM's output can be interpreted probabilistically, but that requires a lot of contortions...) -- Ben - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: Re: Re: [agi] new paper: What Do You Mean by AI?
So you are moving in the right direction. ;-) Pei On 11/18/06, Ben Goertzel [EMAIL PROTECTED] wrote: Agree, too --- that is why I said you want almost everything. However, whenever a design decision is made, you usually consider more about the system's problem-solving ability, and less about the consistency of its theoretical foundation --- of course, you may argue that it don't conflict with your principle, though I'm not that sure. On the contrary, I'm more sensitive to the theoretical consistency of a system than to its practical ability. Pei Hi, Pei ... I think your comments above are more apropos to Webmind than to Novamente. For instance, in my bioinformatics and NLP narrow-AI work I have been extremely impressed with SVM's for supervised categorization. So, in the Webmind philosophy, we would have integrated SVM's into Webmind to serve inside a supervised categorization agent. In Novamente, OTOH, we use MOSES (probabilistic evolutionary learning) for supervised categorization internally (e.g. for finding rules regarding what properties characterize the extension of some concept) ... even though it is not always as accurate as SVM's. The reasons are twofold: a) keeping the overall design as simple as possible b) adherence to the principle of probabilistic semantics (yes, SVM's output can be interpreted probabilistically, but that requires a lot of contortions...) -- Ben - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: Re: Re: Re: [agi] new paper: What Do You Mean by AI?
But of course, practical ability is the foremost thing in my mind when working on AGI design ... my point is just that I will definitely make small sacrifices in the practical ability of particular system components in order to increase the simplicity and principledness of the overall design. Still though, you are right that I remain something of a pragmatic opportunist as an AGI designer, even though I'm a purist re philosophy of mind. Engineering is an opportunistic pursuit, IMO ;=) ben On 11/18/06, Ben Goertzel [EMAIL PROTECTED] wrote: Agree, too --- that is why I said you want almost everything. However, whenever a design decision is made, you usually consider more about the system's problem-solving ability, and less about the consistency of its theoretical foundation --- of course, you may argue that it don't conflict with your principle, though I'm not that sure. On the contrary, I'm more sensitive to the theoretical consistency of a system than to its practical ability. Pei Hi, Pei ... I think your comments above are more apropos to Webmind than to Novamente. For instance, in my bioinformatics and NLP narrow-AI work I have been extremely impressed with SVM's for supervised categorization. So, in the Webmind philosophy, we would have integrated SVM's into Webmind to serve inside a supervised categorization agent. In Novamente, OTOH, we use MOSES (probabilistic evolutionary learning) for supervised categorization internally (e.g. for finding rules regarding what properties characterize the extension of some concept) ... even though it is not always as accurate as SVM's. The reasons are twofold: a) keeping the overall design as simple as possible b) adherence to the principle of probabilistic semantics (yes, SVM's output can be interpreted probabilistically, but that requires a lot of contortions...) -- Ben - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
RE: [agi] new paper: What Do You Mean by AI?
Hi Pei, Just finished reading your Rigid Flexibility book; it's a nice summary of your approach. I can recommend it to anyone interested in AGI: If you agree with Pei's general approach it provides quite a bit a detail; if you disagree, it provides a coherent reference point. Peter -Original Message- From: Pei Wang [mailto:[EMAIL PROTECTED] Sent: Friday, November 17, 2006 8:52 AM To: agi@v2.listbox.com Subject: [agi] new paper: What Do You Mean by AI? Hi, A new paper of mine is put on-line for comment. English corrections are also welcome. You can either post to this mailing list or send me private emails. Thanks in advance. Pei --- TITLE: What Do You Mean by AI? ABSTRACT: Many problems in AI study can be traced back to the confusion of different research goals. In this paper, five typical ways to define AI are clarified, analyzed, and compared. It is argued that though they are all legitimate research goals, they lead the research to very different directions. Furthermore, most of them have trouble to give AI a proper identity. URL: http://nars.wang.googlepages.com/wang.AI_Definitions.pdf - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] new paper: What Do You Mean by AI?
Peter, Thanks! I look forward to the day when you can tell us more about a2i2. :-) Pei On 11/17/06, Peter Voss [EMAIL PROTECTED] wrote: Hi Pei, Just finished reading your Rigid Flexibility book; it's a nice summary of your approach. I can recommend it to anyone interested in AGI: If you agree with Pei's general approach it provides quite a bit a detail; if you disagree, it provides a coherent reference point. Peter -Original Message- From: Pei Wang [mailto:[EMAIL PROTECTED] Sent: Friday, November 17, 2006 8:52 AM To: agi@v2.listbox.com Subject: [agi] new paper: What Do You Mean by AI? Hi, A new paper of mine is put on-line for comment. English corrections are also welcome. You can either post to this mailing list or send me private emails. Thanks in advance. Pei --- TITLE: What Do You Mean by AI? ABSTRACT: Many problems in AI study can be traced back to the confusion of different research goals. In this paper, five typical ways to define AI are clarified, analyzed, and compared. It is argued that though they are all legitimate research goals, they lead the research to very different directions. Furthermore, most of them have trouble to give AI a proper identity. URL: http://nars.wang.googlepages.com/wang.AI_Definitions.pdf - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
RE: [agi] new paper: What Do You Mean by AI?
That'll be when you join our project... ... or buy our product... :) -Original Message- From: Pei Wang [mailto:[EMAIL PROTECTED] Peter, Thanks! I look forward to the day when you can tell us more about a2i2. :-) Pei On 11/17/06, Peter Voss [EMAIL PROTECTED] wrote: Hi Pei, Just finished reading your Rigid Flexibility book; it's a nice summary of your approach. I can recommend it to anyone interested in AGI: If you agree with Pei's general approach it provides quite a bit a detail; if you disagree, it provides a coherent reference point. Peter - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303