Emergence of Awareness and Intelligence

From Self-Organization Wiki
Jump to: navigation, search

(needs to be updated!) (Amardeo Sarma has taken notes)

Working definition

Global (emergent) awareness is the selective amplification and broadcasting of information that is important for the system as a whole. It implies collecting the information sensed by all the components (agents) of the system, and then filtering out of all information that has only local relevance to the agent. The remaining, globally important information must then be made available (broadcasted) to all the agents, so that they can all make use of it to guide their actions.

Intelligence is defined as the ability to solve complex problems. The sense of the word 'complex' is rather subjective, and the topic of intelligence is not treated in this session. The focus is thus on awareness in self-organizing systems.

Examples

  • the "global workspace" in the brain, where subconscious stimuli are made conscious by selective amplification (Stanislas Dehaene and Bernard Baars)
  • swarm conciousness: one or a few birds sense a predator. Their reaction triggers a reaction from the swarm as a whole: all fly away at the same time and in the same direction
  • quorum sensing in SO systems (e.g. via cellular automata rule 184): via local interactions, the system eventually finds out the majority position among its agents
  • emergency handling: some agents detect either a danger (e.g. fire, intrusion), some detect a possible solution (e.. safe exit). This information needs to be made available to all agents collectively
  • awareness of global state in SOS: measure developed by Hermann de Meer

Determinants of Importance

  • relevance to task/goals of the system
  • unexpectedness of information
  • intensity/size: number of agents that independently sensed that information

Conciousness

  • Awareness
    • Similar to knowledge of the global state awareness
    • A selection process is put into place to select the relevant knowledge
      • Competition to select the relevant data
      • Feedback to make decision
      • Aware of most unusual events
  • Representation
    • Abstraction at lower levels

Conciousness in SOS

A pattern of interactions generates conciousness

  • "Global workspace" theory
    • Medium = workspace: many inputs compete depending on the expectedness and the concentration on the task
  • Downward causation / Immergence
    • Interaction between macro and micro
    • micro -> macro: emergence
    • macro -> micro: immergence, i.e. change of perception, selection of micro events, and action on the micro
  • Experience in cellular automata
    • Machine 184: Getting to know the global state (state of the majority) at the local level
    • Competition among local cells to get to this state
  • Flock of birds
    • Example: reaction to predator
    • many stimuli --> many channels --> (collective) decision
    • How is the selection of relevant stimuli made?
      • The importance is related to the context
      • One example could be (it requires action) + (it is important for the group)

Summarizing picture

SOAwareness localpicture.png