Behaviour-based robotics and artificial ethology

The following are quotes from the introductory paragraphs to chapter two of "Behaviour-based Robotics", Ronald C. Arkin, 1998 (MIT Press):

"The possibility of intelligent behaviour is indicated by its manifestation in biological systems. It seems logical then that a suitable starting point for the study of behaviour-based robotics should begin with an overview of biological behaviour. First, animal behaviour defines intelligence. Where intelligence begins and ends is an open-ended question, but we will concede in this text that intelligence can reside in subhuman animals. Our working definition will be that intelligence endows a system (biological or otherwise) with the ability to improve its likelihood of survival within the real world and where appropriate to compete or cooperate successfully with other agents to do so. Second, animal behaviour provides an existence proof that intelligence is achievable. It is not a mystical concept, it is a concrete reality, although a poorly understood phenomenon. Thirdly, the study of animal behaviour can provide models that a roboticist can operationalise within a robotic system. These models may be implemented with high fidelity to their animal couterparts or may serve only as an inspiration for the robotics researcher."

These three points provide the basis for the point of view that animal behaviour has a lot to offer the robotics community (behaviour-based robotics to be specific), and hints at the potential feedback that such work may offer the biologists. Without actually mentioning the term, this description could just as well be applied to artificial (or computational) ethology. The first point I think to be particularly interesting. In my opinion, the working definition of 'intelligence' it introduces is not particularly controversial - however, the implication of the the phrase '...improve its likelyhood of survival...' for cognitive/autonomous robotics is that without some form of actual physical dependancy on the environment (e.g. 'food' - and I hesitantly add, some form of concept of 'life and death' for the agent concerned), intelligence for an artificially created being means nothing (see a related concept in embodiment: organismic embodiment). The second point is one which is generally assumed, but not usually explicitly stated, and something which I think is useful to remind oneself of occasionally. The third point I think is self evident, stated many times, and with plenty of examples in the literature. In fact I think it is the basis for most cognitive robotics work.

The next part of the introduction to chapter two lists two reasons why the robotics community has traditionally resisted the use of the previously mentioned methods of creating artificial agents with 'useful' behaviours (e.g. perceiving and acting in an environment):

"First, the underlying hardware is fundamentally different. Biological systems bring a large amount of evolutionary baggage unnecessary to support intelligent behaviour in their silicon based counterparts. Second, our knowledge of the functioning of biological hardware is often inadequate to support its migration from one system to another. For these and other reasons, many roboticists ignore biological realities and seek purely engineering solutions."

The second point is, I feel, perfectly justified. One only has to consider, for example, the complexity of natural neurons and networks in comparison to the most advanced artificial neural networks which use population-based firing rates, to see that this is true. The first point however, I don't think is necessarily true, especially if one considers that the biological hardware which 'produces' the intelligent behaviour we seek holds many of answers. In this case, an understanding of the 'evolutionary baggage' which produces the biological hardware would be of importance when seeking to understand the intelligent behaviour itself. Or so I think, anyway.

Encephalon #35

I've been slacking recently when it comes to updating the blog... Anyways...

The 35th edition of Encephalon is now up at The Primate Diaries, with 17 of the best neuroscience (and related) posts from the last fortnight. I particularly liked the post from Pure Pedantry with some videos of synapse activation by insertion of AMPA-R into neurons (if I understand correctly, you're much better off reading it for yourself). Fascinating!

On Robots and Psychology

Valentino Braitenberg's Vehicles are often the first lesson in (cognitive) robotics courses as a prime example of the interaction between an agent and its environment, and how complex behaviour does not necissarily imply a complex control architecture (as well as Rodney Brooks' work of course). The fourteen vehicles (or agents) of increasing complexity demonstrate an 'evolution' of behaviours, dependant as much on the environment as on the morphologyand control architecture of the agents themselves. I think the following paragraphs, quoted from the introduction to the book (full reference at end) illustrates how this examination may lead to insight into the biological organisms from which they were inspired (whilst not claiming to 'solve' any of the problems):

I have been dealing for many years with certain structures within animal brains that seemed to be interpretable as pieces of computing machinery because of their simplicity and/or regularity. Much of this work is only interesting if you are yourself involved in it. At times, though, in the back of my mind, while I was counting fibres in the visual ganglia of the fly or synapses in the cerebral cortex of the mouse, I felt knots untie, distinctions dissolve, difficulties disappear, difficulties I had experienced much earlier when I still held my first naive philosophical approach to the problem of the mind. This process of purification has been, over the years, a delightful experience. The text I want you to read is designed to convey some of this to you, if you are prepared to follow me not through a world of real brains but through a toy world that we create together.

We will talk only about machines with very simple internal structures, too simple in fact to be interesting from the point of view of mechanical or electrical engineering. Interest arises, rather, when we look at these machines or vehicles as if they were animals in a natural environment. We will be tempted, then, to use psychological language in describing their behaviour. And yet we know very well that there is nothing in these vehicles that we have not put in ourselves. This will be an interesting educational game.

I like this, as I feel that it represents in some way (albeit more poetically than is generally stated) one of the aims of cognitive robotics as a field: elucidating issues in psychology/neuroscience, via a process (which I have vastly oversimplified here) of model creation on the basis of some biological system, implementation of said model (embodiment in the real world, or simulation thereof), and then evaluation of the resulting system against the original biological system. Hence the emphasis on behaviour - as a means of performing this comparison (since the 'computational substrate' is so obviously different) - which leads to the field of Artificial Ethology.

Ref: Valentino Braitenberg, "Vehicles: experiments in synthetic psychology", 1984

Hammers and Distributed Memory

In “The Feeling of what Happens” (1999), Antonio Damasio describes, among many other things, the distributed nature of memory. The following quote describes an example of how the concept of an object (in this case a hammer) may be represented in the brain (from page 220 of the book):

The brain forms memories in a highly distributed manner. Take, for instance, the memory of a hammer. There is no single place of our brain where we will find an entry with the word hammer followed by a neat dictionary definition of what a hammer is. Instead, as current evidence suggests, there are a number of records in our brain that correspond to different aspects of our past interaction with hammers: their shape, the typical movement with which we use them, the hand shape and the hand motion required to manipulate the hammer, the result of the action, the word that designates it in whatever many languages we know. These records are dormant, dispositional, and implicit, and they are based on separate neural sites located in separate high-order cortices. The separation is imposed by the design of the brain and by the physical nature of our environment. Appreciating the shape of a hammer visually is different from appreciating its shape by touch; the pattern we use to move the hammer cannot be stored in the same cortex that stores the pattern of its movement as we see it; the phonemes with which we make the word hammer cannot be stored in the same place, either. The spatial separation of the records poses no problem, as it turns out, because when all the records are made explicit in image form they are exhibited in only a few sites and are coordinated in time in such a fashion that all the recorded components appear seamlessly integrated.

If I give you the word hammer and ask you to tell me what hammer means, you come up with a workable definition of the thing, without any difficulty, in no time at all. One basis for the definition is the rapid deployment of a number of explicit mental patterns concerning these varied aspects. Although the memory of separate aspects of our interaction with hammers are kept in separate parts of the brain, in dormant fashion, those different parts are coordinated in terms of their circuitries such that the dormant and implicit records can be turned into explicit albeit sketchy images, rapidly and in close temporal proximity. The availability of all those images allows us, in turn, to create a verbal description of the entity and that serves as a base for the definition.

This 'story' of the recall of an abstract concept (by which I mean something which is not explicitly tied to a particular sensory experience) describes a situation where memory, as is generally thought of (recalling objects and events), is fully distributed throughout the brain, and can not be localised in a particular region. This concept has had much supporting empirical evidence found in recent years, including that I've reviewed in last weeks series of posts (which I believe was first proposed by Lashley in the early 20th century).