The Interaction Tasks are a small set of tasks (researchable CSS questions) that will be used to link activity in different CSS application areas and facilitate knowledge exchange across the different CSS approaches and perspectives. Participation in the Interaction Tasks is open to any interested parties in CSIRO, Universities or other institutions. more...

• Emergence
• Human Landscape Interaction
• Memes
• The Continuous/Discrete Interface

CSS Meeting Space

The CSS meeting space is an interactive site that hosts discussions on the Interaction Tasks and supports document sharing. All potential users need to register to access the site.

CSIRO Staff - please register here

Non-CSIRO Users - please note - to register you will need to complete and sign a registration form and fax it to the CSIRO Centre for CSS, as indicated on the form.

Emergence

(Convener - Fabio Boschetti)

"Does anything emerge?" A review on the state of the art in the understanding, modelling and formal description of emergence.

Project Description. Despite the fervor of activities in CSS around the world, the scientific community is still unable to define when a system is complex. In an attempt to provide a working definition of CSS, 'emergence' is often chosen as the most distinguishing feature of a complex system. This may be a circular definition, since the concept of 'emergence' itself is undefined. Also, a number of established researchers question whether emergence actually exists and/or whether the concept is needed at all in the description of Nature.

We see the need to review the state of the art in the understanding, defining, modelling and formal description of emergence. This is motivated by the publication of new approaches and new mathematical tools within the physics and maths community (see below). We believe that the communication of the result of this survey to staff involved in CSS activities can

• provide insights into the science underlying other CSS projects;
• make staff aware of mathematical theories/tools under development in adjacent areas of science;
• help identify a specific aspect(s) of 'emergence' science amenable to investigation in the form of a full project in the coming year.

We plan to analyse and review the following material and further sources arising from it:

• Computational view of emergence; emergence as a hierarchical modelling of a system's internal information processing mechanisms (Jim Crutchfield).
• Evolutionary perspective; emergence arising/modelled by the interaction of simple 'rules' under evolutionary pressure (John Holland).
• Sceptical perspective; emergence as a consequence of our ignorance and inability to model Nature (Cosma Shalizi).
• Effective Field Theory approach: in order to match conditions at different scale boundaries via Renormalisation Groups concepts (Elena Castellani).
• Time Scale Calculus: a tool to reconcile calculus and difference equations with applications to multi scale (time) problems (Martin Bohner).
• Emergence of complexity as a fundamental consequence of an inherently incomplete representation (information processing) of Nature (Anton Zeilinger).
• Emergence in creative thinking and analogy making (Melanie Mitchell).
• Emergence of complexity in basic number theory (Gregory Chaitin).
• Process metaphysics: inability to model emergence as a wrong underlying metaphysical approach (Mark Bickhard).
• Thermodynamic Emergence: emergence of order from apparent noise.

Current development. We have produced a preliminary document, containing the material we collected in our initial literature review. In particular, the document includes mainly 5 sections:

1) a short introduction containing some basic concepts of emergence and two proposed exercises which will drive the discussions at the workshop. This is the part I hope most of you will read even if you can not afford to read the entire document;
2) a brief description of computer 'tools' commonly used to model emergence (or which, according to common opinion, 'display' emergence);
3) a brief section on 'tools/methods' to 'detect' emergence;
4) a longer section containing extracts from several papers which I found particularly insightful. These papers have been selected either because their material is 'different' or more 'in depth' compared to traditional takes on emergence.
5) A couple of appendixes with some definitions of emergence and some information-theoretic measures of complexity which may be useful in our discussion.

Click here to download a PDF version of the current document

Human Landscape Interaction

(Convener - Freeman Cook)

This interaction task will look to foster discussion among the participants on how to incorporate biophysical, social and economic descriptions into complex systems models.

Two topics have been identified for the first year:Identifying spatial and temporal scales that are appropriate to both biophysical, economic and social analysis. Determining the amount of detail that is required in social/economic agents to produce a realistic representation in agent based models and how to develop diagnostic tests for goodness of representation.

Memes

(Convener - Roger Bradbury)

We expect to create a research agenda at the intersection of memes and complex systems that builds on the agent-based modelling approach. A key strength of the this approach is that it allows the many different components of a complex system, from physical to biological to social, to interact naturally. We anticipate a research program that will extend this paradigm to include ideas themselves as agents, using the Darwinian theory of memes. Thus ideas, policies, concepts and beliefs will be allowed to have a life of their own rather than be described as properties of individuals or groups of human beings. This will allow the modelling of a wider range of dynamics of social systems in a new and theoretically challenging way.

The Continuous/Discrete Interface

(Convener - Geoff James)

This topic came up again and again in our attempts to describe, understand, and control systems exhibiting complex behaviour. It has two aspects.

1) Firstly, we need to know how to deal with mixed domains. Some systems seem to be essentially discrete (sensor/actuator networks, robot swarms, immune systems, genetic coding, cellular automata, packet-switched networks, etc.) yet they interact with continuous phenomena (natural environments, chemical gradients, protein folding, propagation media, etc.). How do we formulate problems, understand complex behaviour, and (if we are engineers) develop useful outcomes for mixed systems?

2) Secondly, and more fundamentally, we need to build our knowledge of both discrete and continuous systems, yet discrete systems lack a substantial theoretical basis that we can take as a stepping-off point. How, then, can we carry lessons across? Can differential-equation-based methods for non-linear systems tell us something about discrete systems? What would be the implications on continuous dynamical systems if we found a methodology for designing discrete systems?