week 12: some attempts to understand brain-like architectures

— a quick intro to operating systems

— a parting look at the amygdala

— simple architectures for controlling behavior

— open-ended architectures

a quick intro to operating systems

slide 2

a parting look at the amygdala

slide 3

rethinking innateness

How to push an algorithmic solution through the inter-generational genomic bottleneck? [cf. compilation]

[from Sloman & Chrisley, 2003]
"If we are correct about later evolutionary developments providing high level conceptual, perceptual and meta-management mechanisms that are used both for self-categorisation and other-categorisation (using 'multi-window perception' as in the duck-rabbit picture, and in perception of attentiveness, puzzlement, joy, surprise, etc., in others), then instead of a new-born infant having to work out by some philosophical process of inductive or analogical reasoning or theory construction that there are individuals with minds in the environment, it may be provided genetically with mechanisms designed to use mental concepts in perceiving and thinking about others. This can be useful for predator species, prey species and organisms that interact socially."

slide 4

Aaron Sloman's quest

Thinking about problems and solution mechanisms should be superseded by thinking about problem spaces and classes of architectures.

Cf. Finlay, B. L. (2007) Endless minds most beautiful, Developmental Science 10:30-34.

slide 5

simple architectures for controlling behavior

The protagonist of Valentino Braitenberg's book Vehicles: Experiments in Synthetic Psychology (1984).

hard-wiring hate and love

Hard-wiring aversion and tropism to light.

slide 7

attraction, repulsion, and not much else, in action

slide 8

complexity from simplicity

slide 9

more on vehicles

But can it scale up?

slide 10

scaling up complexity

How does one get from A to B?

(A) (B)

slide 11

incremental learning in a uniform architecture

One approach to building up functional sophistication: Scott Fahlman's cascade correlation algorithm.

This is a simple example of hierarchical abstraction.

slide 12

open-ended architectures

The subsumption architecture:

slide 13

the basic behavior: avoid

slide 14

adding a new function: wander

slide 15

avoid, wander, explore

slide 16

subsumption architectures and brain scaling

Fitts' Law:

Could this be derived also by assuming a generic subsumption architecture?

Is this at all relevant to the brain scaling patterns discovered by Finlay and others?

slide 17

virtual machines (Sloman)

"[...] Finding out what they [brains] do as controllers or as information processors is a very different task from observing physical behaviour, whether internal or external."

"That is because the most important components of an information processor may be components of virtual machines rather than physical machines. Like physical machines, virtual machines do what they do by virtue of the causal interaction of their parts, but such parts are non-physical (by "nonphysical", we do not mean "not physically realised" or "made ultimately of non-physical stuff" but merely "not easily characterised with the vocabulary and methods of the physical sciences"). Compare the notion of a "propaganda machine"."

slide 18

virtual machines

"Entities in virtual machines can include such things as grammars, parsers, decision makers, motive generators, inference engines, knowledge stores, recursive data-structures, rule sets, concepts, plans, and emotional states, rather than molecules, transistors or neurones."

An example of a component of a virtual machine in biology is the niche of a species. A niche is not a geographical location or a physical environment; for an ant, a badger, and a cat may be in the same physical location yet have very different niches, providing different information for them to process, e.g., different affordances such as opportunities, threats and obstacles (Gibson, 1986)."

slide 19

explanatory ontologies in science

"Sometimes scientific progress requires a change in ontology. For example, the discovery that gases are made of previously unknown particles with new kinds of properties (e.g., molecules with mutually repulsive forces) required an extension of the ontology of physics to accommodate the new entities and their properties. In general the deepest advances (both in science and in the conceptual development of an individual) are those that extend our ontologies — for they open up both new classes of questions to pose and new forms of explanations to be investigated. These are not cases where the ontology can be extended simply by defining new concepts in terms of old ones: far more subtle and complex processes are involved, as explained in (Sloman, 1978, chap. 2) and in (Carnap, 1947)."

slide 20

virtual machine ontologies

"A species of ontological layering that is easier to understand than most is found in computing systems where the ontological level of a virtual machine (e.g., a chess-playing machine, a compiler, a theorem prover, an operating system) is implemented on top of an underlying digital electronic machine, a relation often mediated by a hierarchy of intermediate virtual machines."

"This paper is in part about the ontology required for adequate theories concerning the capabilities of biological organisms such as birds, apes and humans [...] A full account of the processes by which the ontologies used by scientists change or grow is beyond the scope of this paper. However, we illustrate the process by describing some features of the ontology required for scientific investigation of intelligent animals and robots, and an application of the ontology in developing an explanatory architecture, H-CogAff."

slide 21

examples of "ontological blindness"

In physics: hypothesizing phlogiston before the role of oxygen in combustion was understood.

In biology: [your example here].

In psychology of vision: focusing on geometrical reconstruction before the importance of affordances was understood.

slide 22

a meta-ontological device: a generative schema for a class of architectures

The task: explore design space.

An example: exploring ecological niches.

slide 23

CogAff (Sloman): the basic building block

"We offer the CogAff schema only as a first draft, very sketchy, starting point, illustrating the more general point that we need a generative schema."

"The CogAff framework permits combinations of mechanisms producing concurrent processes roughly classified as reactive, deliberative, and meta-management (sometimes labelled "reflective") processes."

slide 24

subsuming subsumption

"Not all architectures include mechanisms corresponding to all parts of the grid. Different architectures will have different components and different communication links between components. For instance, some may have only the reactive layer (which may include several different sub-layers, as in most subsumption architectures."

slide 25

the Omega version

"Omega architectures use an information pipeline, with "peephole" perception and action, as opposed to "multi-window" perception and action. The "upward" portion of the pipeline generates possible actions triggered by the sensory input. Selections among options are made at the top and the chosen options are decomposed into low level motor signals on the "downward" pathway."

"People who have not understood the requirement for concurrent hierarchical processing within perceptual and action sub-systems (what we called "multi-window" perception and action) tend to take the Omega structure for granted, though they may propose different sorts of intermediate mechanisms generating options and different sorts of "top-level" decision-making mechanisms."

slide 26

an alarm component

"Some architectures include one or more "alarm mechanisms", i.e., reactive sub-systems with inputs from many parts of the rest of the system and outputs to many parts, capable of triggering global reorganisation of activities, a feature of many emotional processes. Alarm mechanisms may be separate sub-systems running in parallel with the systems they monitor and modulate, or they may be distributed implicitly within the detailed sub-mechanisms, e.g., in conventional programs using very large numbers of tests scattered throughout the code. The former, more modular, type of alarm sub-system may allow more global forms of adaptation and more global kinds of control options when dealing with emergencies, at the cost of architectural complexity."

slide 27

all together now

The H-CogAff Architecture.

"Another way to distinguish Omega-style from true multi-level perception and action would be to require input and output mechanisms to be non-deliberative. On this view, if deliberative mechanisms are involved in the transformation from low-level to high-level input, and from high-level to low-level action, then the Omega architecture best describes that organism. If, however, the low-level input of an organism is transformed into high-level categories by way of non-deliberative, automatic, blind, reactive processes, that are incapable of considering and comparing alternative high-level interpretations of the same data, then that organism can be said to be engaging in true, multi-level perception."

slide 28

empirical support (sort of)

"Organisms with only a subset of the architectural layers will not be capable of having the variety of emotions and other states that are possible according to the CogAff schema. Obviously if insects lack a deliberative layer they will not be able to have emotions (such as regret!) that require "what if" representational capabilities, as most humans can. If human infants lack deliberative mechanisms they too will be unable to have mental states that depend on them. Various kinds of disorders may also be related to different parts of the architecture."

slide 29

empirical support (sort of)

"Although the layers and columns of the CogAff schema need not correspond to anatomically distinct components of an organism, it is consistent with such differentiation. Furthermore, the fact that the layers in a particular organism evolved at different times might make such differentiation likely. It follows that if, as we conjecture, sensory inputs in humans and some other animals are processed concurrently at different levels of abstraction, with information from the different levels transmitted concurrently to different parts of the architecture, which use the information for different tasks, then we can easily explain empirical results that have led some scientists to postulate different perceptual pathways (e.g., Goodale & Milner, 1992), though we would predict more diverse pathways than empirical evidence suggests so far."

slide 30

so, is the amygdala an emotion organ?

"[...] We take it that a system is in one motivational or affective state rather than another primarily because of the role that state plays in mediating between the way the organism takes the world to be and the organism's actions. Thus, an affective or motivational state is a holistic property of a system, not localizable to the state of a "motivational module" or "affective subsystem". This is why we did not include an "emotion box" in either the CogAff schema or the H-CogAff architecture: The aspects of an organism which are responsible for it being in one affective state (e.g., a particular mood or emotional state) rather than another are not, in general, distinct from the total state of the reactive, deliberative and meta-management systems, their control structures, their interactions, etc. In that sense many affective states are "emergent" properties of interactions between mechanisms."

slide 31

the CogAff research framework

  1. Asking questions about an organism: which of the sub-components and which of the links between components does it have, and what difference would it make if the architecture were different in various ways?

  2. Asking similar questions about the ontologies and forms of representations used by different organisms, e.g., what are the affordances they can detect, and how can they use them?

  3. Considering alternative designs for artificial systems, and investigating the pros and cons of including or omitting some of the submechanisms or links between submechanisms;

  4. Illuminating evolutionary investigations by enabling us to identify and analyse possible evolutionary trajectories in design space and in niche space (Sloman, 2000a, 2001).

  5. Challenging and extending the CogAff framework by noticing when useful proposed architectures do not fit the framework.

slide 32

conclusions

"Even if the precise schema we have proposed proves insufficiently general, there will still be a need for something like it as a unifying framework for AI, theoretical psychology and neuroscience."

Comments?