| Both sides previous revisionPrevious revisionNext revision | Previous revision |
| public:t-713-mers:mers-24:methodology [2024/11/05 12:07] – [ConstructiVist AI Methodology] thorisson | public:t-713-mers:mers-24:methodology [2024/11/06 11:20] (current) – [Methodology] thorisson |
|---|
| |
| =====Methodology===== | =====Methodology===== |
| | What it is | The methods - tools and techniques - we use to study a phenomenon. | | | What it is | The methods - tools and techniques - we use to study a phenomenon. \\ The scientific methodology of any field is derived from the prevailing scientific theory/ies in that field. | |
| | \\ Why Scientific Methodology Matters | ... directly determines our progress when studying a phenomenon -- what we do with respect to that phenomenon to figure it out. \\ ... affects how we think about a phenomenon, including our solutions, expectations, and imagination. \\ ... determines the possible scope of outcomes. \\ ... directly influences the shape of our solution - our answers to scientific questions. \\ ... directly determines the speed with which we can make progress when studying a phenomenon. \\ //... is therefore a **primary determinant of scientific progress.** // | | | The essence of methodology | It is always //philosophical// in part, because (a) scientific theories are always rooted in a philosophical context, and (b) they derive from theory in that field, whose bleeding edge is about unanswered questions, which means it is by definition hypothetical, which means it is rooted in a metaphysical context. | |
| | The Main AI Methodology | AI has never had a proper methodology discussion as part of its mainstream scientific discourse. Only 2 or 3 design approaches to AI can be classified as 'methodologies': //BDI// (belief, desire, intention), //subsumption//, //decision theory//. As a result AI inherited the run of the mill CS methodology/ies by default. | | | \\ Why Scientific Methodology Matters | Scientific methdology: \\ ... directly determines what we do with respect to a phenomenon that we are trying to figure out. \\ ... directly affects how we think about a phenomenon, including our solutions, expectations, and imagination. \\ ... defines the possible scope of outcomes. \\ ... directly influences our answers to scientific questions. \\ ... directly determines the speed with which we can make progress when studying a phenomenon. \\ //... is therefore a **primary determinant of scientific progress.** // | |
| | What We Have Studied In This Course | A particular methodology - or //family// of methodologies - emphasizing certain principles over others. \\ It is a //constructivist-inspired, requirements-driven, non-axiomatic// approach. | | | The Main AI Methodology | A proper discussion about methodology has never been a regular part of AI mainstream scientific discourse. \\ Only a handful of approaches to AI R&D can be classified as 'methodologies': //BDI// (belief, desire, intention), //subsumption//, //decision theory//. As a result AI inherited the run of the mill CS methodology/ies by default. | |
| | \\ Constructi//**ON**//ist AI | Methods used to build AI systems by hand. \\ Rely on a third-person view of the phenomenon under study. Methodologies in this category are //allonomic//. \\ Allonomic methodologies are well-suited for classical engineering, where the //model is known//. | | | \\ Constructi//**ON**//ist AI | Methods used to build AI systems by hand. \\ Rely on a third-person view of the phenomenon under study. Methodologies in this category are //allonomic//. \\ Allonomic methodologies are well-suited for classical engineering, where the //model is known//. | |
| | Examples | Virtually all methodologies we have for creating software are methodologies of the allonomic kind (including BDI, Subsumption, software engineering, decision theory, etc.). | | | Examples | Virtually all methodologies we have for creating software are methodologies of the allonomic kind (including BDI, Subsumption, software engineering, decision theory, etc.). | |
| | \\ Constructi//**V**//ist AI | Methods aimed at creating AI systems that autonomously generate, manage, and use their knowledge. \\ Methodologies in this category are //autonomic// (or //constructivist//). \\ Autonomic methodologies are well-suited for science-oriented engineering, where the //model is not known//. | | | \\ Constructi//**V**//ist AI | Methods aimed at creating AI systems that autonomously generate, manage, and use their knowledge. \\ Methodologies in this category are //autonomic// (or //constructivist//). \\ Autonomic methodologies are well-suited for science-oriented engineering, where the //model is not known//. | |
| | Examples | NARS and AERA are the only AI systems known to be built using an autonomic methodology. | | | Examples | NARS and AERA are the only AI systems known to be built using an autonomic methodology. | |
| | | What We Have Studied In This Course | A particular philosophical approach - or //family// of methodologies - emphasizing certain principles over others. \\ It is a //constructivist-inspired, requirements-driven, non-axiomatic// approach. | |
| |
| \\ | \\ |
| =====ConstructiVist AI Methodology ===== | =====ConstructiVist AI Methodology ===== |
| | \\ What it is | A term for labeling a methodology for AGI based on two main assumptions: \\ **(1)** The way knowledge is acquired by systems with general intelligence requires the automatic integration, management, and revision of data in a way that //infuses meaning// into information structures, and \\ **(2)** constructionist approaches do not sufficiently address this, and other issues of key importance for systems with high levels of general intelligence and existential autonomy. || | | \\ What it is | A term for labeling a methodology for AGI based on two main assumptions: \\ **(1)** The way knowledge is acquired by systems with general intelligence requires the automatic integration, management, and revision of data in a way that //infuses meaning// into information structures, and \\ **(2)** constructionist approaches do not sufficiently address this, and other issues of key importance for systems with high levels of general intelligence and existential autonomy. || |
| | | \\ Basic tenet | That an self-programming systems must be able to handle //new// problems in //new// task-environments, and to do so it must be able to create //new// knowledge with //new// goals (and sub-goals), and to do so their architecture must support automatic generation of //meaning//, and that constructionist methodologies do not support the creation of such system architectures. || |
| | Why It's Needed | Assumes that the system acquires the vast majority of its knowledge on its own (except for a small seed) and manages its own GROWTH on its own. Also, it may change its own architecture over time, due to experience and learning. || | | Why It's Needed | Assumes that the system acquires the vast majority of its knowledge on its own (except for a small seed) and manages its own GROWTH on its own. Also, it may change its own architecture over time, due to experience and learning. || |
| | What it's good for | Replacing present methods in AI, by and large, as these will not suffice for addressing the full scope of the phenomenon of intelligence, as seen in nature. || | | What it's good for | Replacing present methods in AI, by and large, as these will not suffice for addressing the full scope of the phenomenon of intelligence, as seen in nature. || |
| | What It Must Do | We are looking for more than a linear increase in the power of our systems to operate reliably, and in a variety of (unforeseen, novel) circumstances. The methodology should help meet that requirement. || | | What It Must Do | We are looking for more than a linear increase in the power of our systems to operate reliably, and in a variety of (unforeseen, novel) circumstances. The methodology should help meet that requirement. || |
| | Basic tenet | That an self-programming systems must be able to handle //new// problems in //new// task-environments, and to do so it must be able to create //new// knowledge with //new// goals (and sub-goals), and to do so their architecture must support automatic generation of //meaning//, and that constructionist methodologies do not support the creation of such system architectures. || | | Roots | || |
| | Roots | ||| | | | Piaget | proposed the //constructivist// view of human knowledge acquisition, which states (roughly speaking) that a cognitive agent (i.e. humans) generate their own knowledge through experience. | |
| | | Piaget | proposed the //constructivist// view of human knowledge acquisition, which states (roughly speaking) that a cognitive agent (i.e. humans) generate their own knowledge through experience. | | | | von Glasersfeld | "...‘empirical teleology’ ... is based on the empirical fact that human subjects abstract ‘efficient’ causal connections from their experience and formulate them as rules which can be projected into the future." [[http://www.univie.ac.at/constructivism/EvG/papers/225.pdf|REF]] \\ | |
| | | von Glasersfeld | "...‘empirical teleology’ ... is based on the empirical fact that human subjects abstract ‘efficient’ causal connections from their experience and formulate them as rules which can be projected into the future." [[http://www.univie.ac.at/constructivism/EvG/papers/225.pdf|REF]] \\ | | | | \\ NARS | Non-Axiomatic Reasoning System [[https://sites.google.com/site/narswang/|REF]] \\ //“If the existing domain-specific AI techniques are seen as tools, each of which is designed to solve a special problem, then to get a general-purpose intelligent system, it is not enough to put these tools into a toolbox. What we need here is a hand. To build an integrated system that is self-consistent, it is crucial to build the system around a general and flexible core, as the hand that uses the tools [assuming] different forms and shapes.”// -- P. Wang, 2004 | |
| | | \\ NARS | Non-Axiomatic Reasoning System [[https://sites.google.com/site/narswang/|REF]] \\ //“If the existing domain-specific AI techniques are seen as tools, each of which is designed to solve a special problem, then to get a general-purpose intelligent system, it is not enough to put these tools into a toolbox. What we need here is a hand. To build an integrated system that is self-consistent, it is crucial to build the system around a general and flexible core, as the hand that uses the tools [assuming] different forms and shapes.”// -- P. Wang, 2004 | | | Limitations | As a young methodology, very little hard data is available to its effectiveness. \\ What does exist, however, is more promising than constructionist methodologies for achieving GMI. || |
| | Limitations | As a young methodology very little hard data is available to its effectiveness. What does exist, however, is more promising than constructionist methodologies for achieving GMI. || | |
| |
| \\ | \\ |
