public:t_720_atai:atai-18:lecture_notes_methodologies

Table of Contents

T-720-ATAI-2018

T-720-ATAI-2018 Main
Links to Lecture Notes

T-720-ATAI-2018

Lecture Notes, W8: AI Methodologies

Complex Systems

Kind of System	What it Consists of	Theory	Methodology
Static system	Elements of the system do not interact, or interact slowly. Example: Mountains.	Depends on the domain.	Ditto
Dynamic system	Elements of the system interact.	Depends on the domain.	Ditto
Simple system	Few interacting parts	Mechanics	Observation of operation, experimentation, parts analysis. Analytic analysis.
Complex uniform system	Vast amounts of identical elements interacting	Thermodynamics	Statistics. Mathematical models and simulation.
Complex heterogeneous system	Multiple unique elements interacting	–missing!–	Agent-based models and simulation

Minds Are Complex Systems

So You Want To Create Human-Level AI?	If we want to make smarter machines we should begin by isolating the requirements. Requirements for AGI were already covered - that's a long list (but doable, we hope!). Are there other concerns? Yes: We need to pick an appropriate methodology for this task.
Mind: What Kind of System?	Mind contains a large number of processes (large size), of a wide variety implementing a variety of complex functions (heterogeneity), that work closely together in an integrated manner to implement global emergence (dense coupling); it is a heterogeneous large densely coupled system - a HeLD.
HeLDs	HeLDs are found everywhere in nature: Ecosystem, forests, societies, traffic, economies, and yes - minds. In fact, minds are a special kind of self-governed HeLD.
How to Study HeLDs	HeLDs require special consideration of methodology - good old fashioned run-of-the-mill reductionism may not suffice.

Methodology: What It Is

What it is	The methods (tools and techniques) we use to study a phenomenon.
Examples	- Comparative experiments (for the answers we want Nature to ultimately give). - Telescopes (for things far away). - Microscopes (for all things smaller than the human eye can see unaided). - Simulations (for complex interconnected systems that are hard to untangle).
Why it's important	Methodology directly determines how we may study a phenomenon – what we can do with respect to that phenomenon when scrutinizing it. Methodology affects how we think about a phenomenon, including our solutions, expectations, and imagination. Methodology determines the possible scope of outcomes. Methodology directly influences the shape of our solution - our answers to scientific questions. Methodology directly determines the speed with which we can make progress when studying a phenomenon. Methodology is therefore a primary determinant of scientific progress.
The main AI methodology	AI never really had a proper methodology discussion as part of its mainstream scientific discourse. Only 2 or 3 approaches to AI can be properly called 'methodologies': BDI (belief, desire, intention), subsumption, decision theory. As a result AI inherited the run of the mill CS methodology/ies by default.
Constructionist AI	Methods used to build AI systems by hand.
Constructivist AI	Methods aimed at creating AI systems that autonomously generate, manage, and use their knowledge.

How to Study HeLDs Scientifically

Reductionism	The method of isolating parts of a complex phenomenon or system in order to simplify and speed up our understanding of it. See also Reductionism on Wikipedia.
Occam's Razor	Key principle of reductionism. See also Occam's Razor.
HeLD	Cannot be studied by the standard application of reductionism/Occam's Razor, because the emergent properties are lost. Instead, corollaries of the system – while ensuring some commonality to the original system in toto – must be studied to gain insights into the target system.
Agent & Environment	We try to characterize the agent and its task-environment as two interacting complex systems. If we keep the task-environment constant, the remaining system to study is the agent and its controller.

How to tease apart HeLDs.

Relationship between system, its task-environment, and the world. Task-environments will always inherit the “laws” of the world; the world puts constraints on the state-space of the task-environment.

Constructionist Methodologies: Traditional CS Software Development Methods

What it is	A constructionist methodology requires an intelligent designer that manually (or via scripts) arranges selected components that together makes up a system of parts (read: architecture) that can act in particular ways. Examples: automobiles, telephone networks, computers, operating systems, the Internet, mobile phones, apps, etc.
Why it's important	Virtually all methodologies we have for creating software are of this kind.
Fundamental CS methodology	On the theory side, for the most part mathematical methodologies (not natural science). On the practical side, hand-coding programs and manual invention and implementation of algorithms. Systems creation in CS is “co-owned” by the field of engineering.
The main methodology/ies in CS	Constructionist.

Constructionist AI

Constructionist AI	Methodology for building cognitive agents that relies primarily on Constructionist methodologies.
What it is	Refers to AI system development methodologies that require an intelligent designer – the software programmer as “construction worker”.
Why it's important	All traditional software development methodologies, and by extension all traditional AI methodologies, are constructionist methodologies.
What it's good for	Works well for constructing controllers of Closed Problems where (a) the Solution Space can be defined fully or largely before the controller is constructed, (b) there exist clearly definable Goal hierarchies and measurements that, when used, fully implement the main purpose of the AI system, and ( c) the Task assigned to the controller will not change throughout its lifetime (i.e. the controller does not have to generate novel sub-Goals).
Key Implementation Method	Hand-coding using programming language and methods created to be used by human-level intelligences.

Example Constructionist Methodology: Subsumption Architecture

Augmented Finite State Machines (AFSMs)	Finite State Machines, augmented with timers.
Modules (FSMs) have internal state	The internal state includes: * the clock * the inputs (no history) * the (current) output (no history) * may include “activation level”
External environment constists of connections (“wires”)	* Input * Inhibitor * Suppressor * Reset * Output
Augmented Finite State Machine (AFSM) with connections
	Suppressor: Replaces the input to the module Inhibitor: Stops the output for a given period Reset: Initialization puts the module in its original state
Augmentation	The finite state machines are augmented with timers. The time is fixed for each I or R, per module.
Timers	Timers enable modules to behave autonomously based on a (relative) time
The AFSMs are arranged in “layers”	Layers separate functional parts of the architecture from each other


Example subsumption architecture with layers.

Key Limitations of Constructionist Methodologies

Static	System components that are fairly static. Manual construction limits the complexity that can be built into each component.
Size	The sheer number of components that can form a single architecture is limited by what a designer or team can handle.
Scaling	The components and their interconnections in the architecture are managed by algorithms that are hand-crafted themselves, and thus also of limited flexibility.
Result	Together these three problems remove hopes of autonomous architectural adaptation and system growth.
Conclusion	In the context of artificial intelligences that can handle highly *novel* Tasks, Problems, Situations, Environments and Worlds, no constructionist methodology will suffice.
Key Problem	Reliance on hand-coding using programming methods requiring human-level intelligence. So the system cannot program itself. Another way to say it: Strong requirement of an outside designer.
Contrast with	Constructivist AI

Constructivist AI Methodology (CAIM)

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.
Why it's important	It is the first and only attempt so far at explicitly proposing an alternative to current methodologies and prevailing paradigm, used throughout AI and computer science.
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.
Basic tenet	That an AGI 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	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.” REF CAIM was developed in tandem with this architecture/architectural blueprint.
Architectures built using CAIM	AERA	Autocatalytic, Endogenous, Reflective Architecture REF Built before CAIM emerged, but based on many of the assumptions consolidated in CAIM.
	NARS	Non-Axiomatic Reasoning System 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.” – 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 AGI.

Constructivist AI

Foundation	Constructivist AI is concerned with the operational characteristics that the system we aim to build – the architecture – must have.
Behavioral Characteristics	Refer back to the requirements for AGI systems; it must be able to: - handle novel task-environments. - handle a wide range of task-environments (in the same system, and be able to switch / mix-and-match. - transfer knowledge between task-environmets. - perform reasoning: induction, deduction and abduction. - handle realtime, dynamic worlds. - introspect. - …. and more.
Constructivist AI: No particular architecture	Constructivist AI does not rest on, and does not need to rest on, assumptions about the particular kind of architecture that exists in the human and animal mind. We assume that many kinds of architectures can achieve the above AGI requirements.

Examples of Task-Environments Targeted by Constructivist AI

Diversity	Earth offers great diversity. This is in large part why intelligence is even needed at all.
	Desert
	Ocean floor
	Air
	Interplanetary travel
The Same System at the Same Time	These task-environments should be handled by a single system at a single period in time, without the designers coming anywhere close.
Baby Machines	While the mechanisms constituting an autonomous learning “baby” machine may not be complex compared to a “fully grown” cognitive system, they are likely to result in what nevertheless will seem large in comparison to the AI systems built today, though this perceived size may stem from the complexity of the mechanisms and their interactions, rather than the sheer number of lines of code.

Architectural Principles of AGI Systems / CAIM

Self-Construction	It is assumed that a system must amass the vast majority of its knowledge autonomously. This is partly due to the fact that it is (practically) impossible for any human or team(s) of humans to construct by hand the knowledge needed for an AGI system, and even if this were possible it would still leave unanswered the question of how the system will acquire knowledge of truly novel things, which we consider a fundamental requirement for a system to be called an AGI system.
Semiotic Opaqueness	No communication between two agents / components in a system can take place unless they share a common language, or encoding-decoding principles. Without this they are semantically opaque to each other. Without communication, no coordination can take place.
Systems Engineering	Due to the complexity of building a large system (picture, e.g. an airplane), a clear and concise bookkeeping of each part, and which parts it interacts with, must be kept so as to ensure the holistic operation of the resulting system. In a (cognitively) growing system in a dynamic world, where the system is auto-generating models of the phenomena that it sees, each which must be tightly integrated yet easily manipulatable and clearly separable, the system must itself ensure the semiotic transparency of its constituents parts. This can only be achieved by automatic mechanisms residing in the system itself, it cannot be ensured manually by a human engineer, or even a large team of them.
Self-Modeling	To enable cognitive growth, in which the cognitive functions themselves improve with training, can only be supported by a self-modifying mechanism based on self-modeling. If there is no model of self there can be no targeted improvement of existing mechanisms.
Self-Programming	The system must be able to invent, inspect, compare, integrate, and evaluate architectural structures, in part or in whole.
Pan-Architectural Pattern Matching	To enable autonomous holistic integration the architecture must be capable of comparing (copies of) itself to parts of itself, in part or in whole, whether the comparison is contrasting structure, the effects of time, or some other aspect or characteristics of the architecture. To decide, for instance, if a new attention mechanism is better than the old one, various forms of comparison must be possible.
The “Golden Screw”	An architecture meeting all of the above principles is not likely to be “based on a key principle” or even two – it is very likely to involve a whole set of new and fundamentally foreign principles that make their realization possible!

Some Key Requirements For a Constructivist AGI Architecture

Tight Integration	A general-purpose system must tightly and finely coordinate a host of skills, including their acquisition, transitions between skills at runtime, how to combine two or more skills, and transfer of learning between them over time at many levels of temporal and topical detail.
Holistic Integration	The architecture of an AGI cannot be developed in a way where each of the key requirements (see above) is addressed in isolation, or semi-isolation, due to the resulting system's whole-part semiotic opaqueness: When a system learns new things, to see whether it has learned it before, and use it to improve its understanding, it must relate the new knowledge to its old knowledge, something we call *integration. The same mechanisms needed for integration also enable transfer knowledge; it is these same mechanisms that (in humans) are responsible for what is known as “negative transfer of training”, where a priorly learned skill makes it harder* to learn something new (this happens in humans when the new task is almost like the old one, but deviates on some points. The more critical these points are in mastering the skill, the worse the negative transfer of training.
Transversal Functions	The system must have pan-architectural characteristics that enable it to operate consistently as a whole, to be highly adaptive (yet robust) in its own operation across the board, including metacognitive abilities. Some functions likely to be needed to achieve this include attention, learning, analogy-making capabilities, and self-inspection.
Time	Ignoring (general) temporal constraints is not an option if we want AGI. Move over Turing! Time is a semantic property, and the system must be able to understand – and be able to learn to understand – time as a real-world phenomenon in relation to its own skills and architectural operation. Time is everywhere, and is different from other resources in that there is a global clock which cannot, for many task-environments, be turned backwards. Energy must also be addressed, but may not be as fundamentally detrimental to ignore as time while we are in the early stages of exploring methods for developing auto-catalytic knowledge acquisition and cognitive growth mechanisms.
Large Aarchitecture	An architecture that is considerably more complex than systems being built in most AI labs today is likely unavoidable. In a complex architecture the issue of concurrency of processes must be addressed, a problem that has not yet been sufficiently resolved in present software and hardware. This scaling problem cannot be addressed by the usual “we’ll wait for Moore’s law to catch up” because the issue does not primarily revolve around speed of execution but rather around the nature of the architectural principles of the system and their runtime operation.
Predictable Robustness	The system must have a robustness in light of all kinds of task-environment and embodiment perturbations, otherwise no reliable plans can be made, and thus no reliable execution of tasks can ever be reached, no matter how powerful the learning capacity.
Graceful Degradation	Part of the robustness requirement is that the system be constructed in a way as to minimize potential for catastrophic failure. A programmer can forget to delimit a command in a compiled program and the whole application crashes; this kind of brittleness is not an option for cognitive systems that operate in stochastic environments, where perturbations can come in any form at any time.
Time is Integrated	Time must be a tightly integrated phenomenon in any AGI architecture - managing and understanding time cannot be “retrofitted” to a complex architecture!

What Are Methodologies Good For?

Applying a Methodology	results in a family of architectures: The methodology “allows” (“sets the stage”) for what should and may be included when we design our architecture. The methodology is the “tool for thinking” about a design space. (Contrast with requirements, which describe the goals and constraints (negative goals)).
Following a Methodology	results in a particular architecture.
CAIM Relies on Models	CAIM takes Conant & Ashby's proof (that every good controller of a system is a model of that system - the Good X Theorem) seriously, putting models at its center This stance was prevalent in the early days of AI (first two decades) but fell into disfavor due to behaviorism (in psychology and AI).
Example	The Auto-Catalytic Endogenous Reflective Architecture - AERA - is the only architecture to result directly from the application of CAIM. It is model-based and model-driven (in an even-driven way: The models left-hand terms are matched to situations to determine their relevance at any point in time, when they match their right-hand term is injected into memory - more on this below).
In Other Words	AERA Models are a way to represent knowledge. But what are models, really, and what might they look like in this context?

2018©K. R. Thórisson

EOF