CSCE475/875 Multiagent Systems

Handout 19: Collaborative Topic Summary Assignment 6: Q & A

November 13, 2009

>>  Several questions were selected from Assignment 6.  Here is my response to these questions.

Q1.  What level, of an n-level agent, would be equivalent of human intelligence? What kind of storage structure would be used to simulate human intelligence?

Response:  Human intelligence is limitless, from one perspective.  If we say that what would be considered to exhibit human intelligence, then as easy as 1-level agent can suffice.  Remember the ELIZA system that I talked about—the one that passed the Turing Test?  That system was simply a 2-level agent, modeling what its counterpart was saying. 

There is another issue here: the higher the n, the more complex the system is.  So, we should also consider whether there is a need for n, or n-1, or n-2, etc.  Practicality has to be factored in when designing a MAS. 

In terms of storage structure, some have argued that a hierarchical structure with IsA-links and HasA-links is sufficient to represent states of human reasoning or represent domains of human knwoledge.  Note that the question here used the word “simulate” human intelligence.  To simulate human intelligence, I do not what storage structure would be. 

Q2.  Are some learning techniques preferable in dynamic or dishonest agent environments where unlearning may become necessary?

Response:  Let’s clarify something first.  It seems that the assumption here is that in dynamic or dishonest agent environments, unlearning might be needed.  I see learning differently.  And dynamic or dishonest agents are two different scenarios.  If there are agents who are dishonest, if I am learning about them, then I do not have to unlearn.  Sooner or later, I will learn that  those agents’ information cannot be trusted as I find out the utility of the actions that I have carried out based on information is low, for example.  Now, if you are referring to information that I obtained from them, and now I need to discard that information and thus this is considered “unlearning”, then a fault-diagnosis-based approach can be used.  Find the faulty information component that caused a negative utility result, for example, and then remove that information component.  This is actually related to Truth Maintenance Systems (TMSs) in AI.  Very interesting area.    Now, with “dynamic environments”, well, agents are supposed to behave in such environments, and many learning techniques such as the Q-learning-based ones are exactly suitable for such environments

Q3.  Two similar questions: (a) In Market Environment, is it possible to have agents with more than 2-level?  (b) In market modeling, can an agent model multiple levels of agents? Are there any levels where the agent models others as at least the same level as itself.

Response:  Let’s see this first:  A 3-level agent would be what?  A 1-level agent models 0-level agents.  A 2-level agent models 1-level agents.  A 3-level agent would have to model 2-level agents.  Would that be useful?  Think about the need for 2-level.  In this case, a 3-level agent would model 2-level agents.  But I 2-level agents only model 1-level agents, should the 3-level agent even bother to model those agents that way? 

It will be difficult for, for example, a 2-level agent to model another agent as a 2-level agent.  Why?  There would be an infinite loop.  Think about paper-scissor-rock the game.  Let’s say, I model you as a 0-level player.  That is, you simply choose a shape without considering what I would pick.  So, then, I can say “You are most likely to choose paper.” And then I can follow with “scissor”.  But, if I start to model you as a player who also models me, then I start to get into an infinite loop as I cannot arrive at a stable choice.  The same thing applies to the market environment n-level agents.  Note that in the market environment, it is okay to have ALL agents to be 2-level agents, for example.  But it will be difficult for a 2-level agent to model another agent (even if that agent is a 2-level agent) as a 2-level agent. 

Q4.  In artificial intelligence, there exists a notion of state which includes a description of the current environmental conditions, the known contents of that environment and the current value of any variables that are internal with respect to a particular agent.  As the environment and other contributors of state become more complex, the state representation becomes more complex.  Furthermore, the number of states and the complexity of relationships among states can increase.  Is there a better way to represent state?  In real life, we often only consider all that might be involved in a particular state in time when we take care to recognize what's actually going on in our environment.  We ignore many things; for example, one state might be considered as a span of time rather than a particular instant.  We may ignore passing cars on a highway, or apply a general concept to represent the system that includes the highway and the cars moving along it.  How might a state be represented flexibly enough to represent functional information or particular general models, or exact state quantities? If the knowledge representation is different between agents, do they have to learn to understand the semantic meaning of the messages first when they are communicating for learning?

Response:  This is a long question.  In AI, state representation has always been a main research area.  The frame problem, the ramification problem, and the representation problem are the three main problems in state representation.  

Yes, in real-life, we have to make assumptions in order to function.  And yes, we do ignore many things in order to make decisions, to act, and to reason.  

To allow for reasoning with functional information, particular general models, or exact state quantities, we usually use a state hierarchy.  Sometimes, we use ISA links, HAS links, causal links, co-occurrence links, etc., to connect states.  We also use belief-desire-intention (BDI) to represent our confidence/perception of states.  

For the last question, yes, if the knowledge representation is different, then either there is an ontology that maps between the two representations, or the agents do have to learn to do that.  There has been some work in ontology learning in multiagent systems.  Quite interesting.