|
cas - complex adaptive system
s/b - signal/boundary
dgs - dynamic generated system
p.1 Ecosystems, governments, biological cells, markets, and complex adaptive systems in general are characterized by intricate hierarchal arrangements of boundaries and signals... It is the stance of this book that widely applicable answers to questions about steering complex adaptive systems can be attained only by studying the origin and the coevolution of signal/boundary hierarchies, much as the intricacies of ecosystems can be understood only by studying the origin and the coevolution of species. [JLJ - can a book take a stance?]
p.2 The critical role of mechanisms is nicely illustrated by Jacob Bjerknes' discovery of fronts as a weather-generating mechanism... Bjerknes' discovery suggested where to look for new relevant data, and how to make predictions conditional on the data. Similarly, the objective here is to find mechanisms that suggest where to look for new data that explain the development of signal/boundary hierarchies. The book's ultimate goal is to tie these mechanisms into a single overarching framework that suggests ways to steer complex adaptive systems by modifying signal/boundary hierarchies.
p.3 Consider the monkey's cry. Clearly it is a signal, but what are the boundaries? We must look more closely at the monkey's environment to see the relevant boundaries.
p.5-6 When there are many reproducing individuals, coevolution puts a premium on interactions between individuals - predator and prey, symbiosis, and the like. Individual-to-individual interaction, in turn, puts emphasis on signals to mediate the interactions. Ever-increasing complexity follows, stemming directly from the evolving boundaries and signals.
p.6 At a fundamental level, thought itself depends on our penchant for seeing the world in terms of the bounded shapes we call objects. To see an object, we must organize the perpetually novel, confusing array of light signals striking the eye into familiar, bounded shapes. Once we organize our sensory input into objects, we can go on to parse the world about us in increasingly sophisticated ways.
p.11 this book... its objective is to better understand the origins and effects of signal/boundary interactions, wherever they may occur. [JLJ - Great. We will look at them with respect to game theory.]
p.20 Every time we examine a signal/boundary system closely, we see coevolution as a pervasive feature... important signal/boundary interactions are involved.
p.20 The mechanisms and interactions falling under the broad categories diversity, recirculation, niche and hierarchy, and coevolution are central to understanding a wide range of large-scale and small-scale signal/boundary systems. If we can fit these categories within an overarching framework, it will further our quest for a general understanding of signal/boundary phenomena.
p.24-25 Chapter by chapter, this book introduces a sequence of concepts that can be melded into an appropriate framework:
classifier systems for defining signal-processing programs
tags for directing signals
tagged urns for defining semi-permeable boundaries
genetic algorithms to provide for the adaptation and coevolution of agents
dynamic generated systems that provide a "grammar" for the framework and bring mathematical tools to bear.
The book's major task is to fit these concepts together to form a well-defined framework... each chapter illustrates the new concepts by using biological cells as a running example.
p.30 How do finely tuned boundaries and signals come into being? It is clear that coevolution, in one sense or another, is involved in every case. Indeed, this book's major thesis is that the key to understanding complex signal/boundary interactions is an understanding of their progressive coevolution.
p.35-36 When we face difficult scientific questions that cannot be answered by inspection or by trial and error, formal theory has an essential role.
p.37 The first step toward a signal/boundary theory, then, is to phrase the questions we would like to address in ways that suggest premises for a deductive system.
p.38 Anatol Rapoport, one of the founders of mathematical biology, long ago pointed out that you cannot learn the rules of chess by keeping only the statistics of observed moves (Rapoport 1960). We confront the same difficulty when using statistics to study signal/boundary interactions. The interactions are just too complex (nonlinear) to allow theory to be built with the linear techniques of statistics.
p.40 A good model or theory suggests controlled experiments to confirm or disconfirm the hypothesis it poses. More than that, models and theories have made, and do make, verifiable predictions in situations not previously encountered. A model based on mechanisms can be quite explicit in its suggestions about where to look for new possibilities.
p.41 Computer-based models have become a major tool for investigating complex adaptive systems because they can handle conditional "IF this happens, THEN take this action" interactions. Such interactions pose difficulties for traditional equation-based models because, in mathematical terms, they are nonlinear (non-additive).
p.43 Typically, exploratory models start with a designated set of mechanisms... with the objective of finding out what can happen when these mechanisms interact.
p.44 Exploratory models show the adequacy or the inadequacy of a set of mechanisms for generating a certain range of signal/boundary observations.... Exploratory models concerned with signal/boundary interactions are the core of this book.
p.45 A model's clarity and generality depend directly on how much detail has been set aside.... One way to construct an overarching signal/boundary model is to eliminate features not held in common by different signal/boundary systems. By eliminating these details, the model deals with recurrent phenomena - common features that occur and reoccur in the dynamics of different signal/boundary systems. Recurrent phenomena supply building blocks for the model.
p.46 what is detail for one theory may prove to be critical for another. As was emphasized earlier, detail depends on the question being posed.
p.51-55 In constructing a general signal/boundary theory, building blocks have high priority because they offer a way to approach the formation of complex signals and boundaries from simple elements... the building block approach to complexity provides us with a set of four specific requirements for a constructive signal/boundary theory. [JLJ - bullet notation below added for clarity]
- Requirement 1 Signals and boundaries, and all things employing them, should be defined with the help of a formal grammar that specifies allowable combinations of building blocks... A small set of generators... is transformed by a small set of rules... into a large array of objects of interest... the generators and the generating rules must be chosen so that objects generated relate to the questions we hope to answer... For signal/boundary questions, the generators must generate boundaries, signals and, particularly, the signal/boundary conglomerates that serve as agents...
- Requirement 2 Each generator used by the signal/boundary grammar should have a location in an underlying geometry, and combinations of generators should be mobile within that geometry...
- Requirement 3 The grammar should be capable of generating programmable agents - bounded conglomerates that can execute arbitrary signal-processing programs...
- Requirement 4 The signal/boundary grammar must provide for reproduction by collecting resources, whereby an agent-conglomerate reproduces by collecting copies of the generators that define its structure...
p.58 In a complex adaptive system, equilibria are rare and temporary. Adaptation by recombination of "building blocks" is a continuing process at all levels, and coevolution clearly has a central role in this continuing adaptation.
p.59 Because the agents incorporate adaptive mechanisms, the systems continue to innovate. In other words, the mechanisms of change are primarily mechanisms of exploration rather than exploitation.
p.59 The immediate reaction of an agent to the signals from other agents is conditional, taking a condition/action form... At any specific time, different agents will have different condition/action combinations, resulting in the complex interactions that typify a complex adaptive system. If the situation is changed slightly, the agents respond differently.
p.60 evolutionary adaptation usually proceeds by introducing new agents formed by recombination of "building blocks" present in extant agents.
p.60 anticipation doesn't require consciousness, even though anticipation is usually discussed in the context of consciousness... In general, anticipation appears whenever an agent, conscious or otherwise, acquires an internal model of its environment. In a cas, the environment includes other agents, so the internal model usually includes models of other agents. Then, conscious-like actions begin to appear.
p.62 It is important for agent-based signal/boundary studies to keep longer-term, coevolutionary changes at center stage.
p.71 An agent is situated when it is placed in an environment that consists of other agents and "inanimate" objects such as resources and patches of resources. To interact with this environment, the agent must have a way of detecting local features of the environment and a way of modifying that environment... For cas signal-processing agents, we will invoke a set of detectors for determining the changing features of the surrounding environment... The detector array serves as an interface that transforms properties of the environment into a binary signal string that serves as input to the situated agent.
p.71-72 Each effector of a situated agent defined by a classifier system is activated by an appropriate signal on the agent's signal list. Many effectors can be active simultaneously when the appropriate signals are on the signal list. Though the signals themselves cannot conflict, the actions of different effectors can conflict... If simultaneous signals try to activate both effectors, the ensuing conflict must be resolved... A more interesting approach... uses a notion of signal strength - the effector receiving the stronger signal is the one activated.
p.88 In more technical terms, in response to changing conditions in the niche, they change the phenotypic activities that define their niche behavior.
p.99 Chapter 4 made the point that in order to understand signal/boundary systems we must consider the formation of boundaries and signals, not just their existence. In complex adaptive systems in general, and in signal/boundary systems in particular, adaptive mechanisms mediate the formation of new structures
p.100 In the formation of new boundaries, signals influence the structure of new boundaries at least as often as boundaries influence the formation of new signals. Changes in both signals and boundaries occur on many different time scales.
p.108 the "machine code" of a contemporary computer chip usually involves 32 or 64 basic instructions, and a program is simply a sequence of instructions. [JLJ - either badly worded or misunderstood. Such generalizations concerning instruction sets should be more specific as to what you are actually counting as a 'basic instruction'.]
p.109 Both innovation and diversity result from combining familiar building blocks in new ways... Innovation through recombination of building blocks leads to sustainable diversity.
p.110-111 by extracting building blocks and examining their combinations we acquire a way to understand a wide spectrum of complex adaptive systems... Building blocks are often not at all evident when we first look at some complex adaptive systems. This is so even though we have observed thousands of different structures that, later, we find be constructed from a set of simple building blocks... a search for barriers or filters in a cas often guides the search for building blocks. Moreover, in a cas the building blocks of interest are processors - they carry out conditional actions... It is conditional action, based on the context provided by the other building blocks, that produces complex behavior from simple building blocks
p.111 A building block is a generator if it is immutable over the span of its existence... If a generator is removed from a structure, the structure is changed, typically to a structure with a very different function or interpretation... A building block is a conglomerate if it can grow and "fission" into additional conglomerations related to the "parent" conglomerate... In most cases, a conglomerate can be defined in terms of an interconnected set of generators... The pattern that defines the conglomerate usually doesn't persist as long as its component generators. However, a conglomerate isn't of much interest unless it persists long enough to serve as a building block for more complex structures.
p.112 This distinction between generators and conglomerates is often ignored by researchers, but it can be important.
p.112 Because a conglomerate has parts, it can become a source of new conglomerates through recombination with other extant conglomerates. When the new conglomerates act as building blocks for new adaptations, the process can be repeated, producing still further adaptations, resulting in a system that is perpetually changing. In such circumstance, the conglomerate is best interpreted as a persistent pattern imposed on a flow of building blocks.
p.113 Conglomerates, as patterns imposed on flows, can be modified to augment properties such as self-repair and homeostasis (Ashby 1952), thereby further increasing their persistence... To repeat an important point: Conglomerates can usually be described in terms of changing combinations of building blocks drawn from the next lower level of description... Accordingly, conglomerates will be discussed in terms of their constituent building blocks wherever that is possible.
p.114 In complex adaptive systems, emergent properties often occur when coevolving signals and boundaries generate new levels of organization. Newer signals and boundaries can then emerge from combinations of building blocks at this new level of organization.
p.124-125 Darwinian selection... There are three requirements:
- There must be a population so that the rules can compete.
- There must be a comparison of rules that results in a rating that indicates the rule's performance within the population.
- There must be a selection process, so that rules that are successful in the competition can be favored.
p.130 In particular, rules that act early, setting the stage for the later reservoir-filling action, become stronger.
In this way strength and fitness are tied to each other because chains that lead to reservoir filling actions are strengthened and play bigger roles in determining the agent's behavior. Successful chains fill the reservoirs more rapidly, allowing the agent to replicate more rapidly. In Darwinian terms, such an agent is more fit and contributes more to the evolution of the signal/boundary system.
p.148 One broad conjecture about hierarchies of enclosure is that they increase the rate of production of selectively advantageous reactants.
p.191 The rule-based cas approach used here emphasizes the generation of sequenced utterances, or gestures, that (sometimes) steer other agents... Generally the utterances are in response to a salient environmental situation... For a rule-based approach to be realistic, an agent must have many simultaneously active rules that emit internal coordinating signals. This simultaneous activity is roughly the counterpart of the simultaneous firing of assemblies or neurons in the central nervous system (Hebb 1949).
p.191-192 In the LoC approach, the overall activity of the agent is conditioned on three factors:
A desire or need.. The current environmental situation... An anticipation... An utterance issued at any given point in time is typically conditioned upon signals of all three types. The agent's objective, in each situation, is to use a sequence of utterances to cause a desired change in its environment.
When an anticipated outcome doesn't happen, search and learning are activated.
p.214 The expectations are so strong that it is still possible to guess the meaning ni wtretin Esngilsh wneh teh oderr fo het leertts si sbcrlmabed.
|
