CSCE 475/875
Handout 6:
Learning and Communication
September 13,
2011
This
handout is based on Chapter 6 of G.
Weiss, (Ed.), Multiagent Systems: A Modern Approach to Distributed
Artificial Intelligence, MIT Press, 1999.
What
is learning?
The acquisition of new knowledge and motor and cognitive skills
and the incorporation of the acquired knowledge and skills in future system
activities, provided that this acquisition and incorporation is conducted by
the system itself and leads to an improvement in its performance.
Differencing Features
The degree of decentralization.
Concerning distributedness and parallelism. One extreme is that a single agent carries
out all learning activities sequentially.
The other extreme is that the learning activities are distributed over
and parallelized through all agents in a MAS.
Interaction-specific features.
There is a number of features that can be applied to classifying the
interactions required for realizing a decentralized learning process.
·
the level of interaction (ranging from pure observation over simple signal passing and
sophisticated information exchange to complex dialogues and negotiations),
·
the persistence of interaction (ranging from short-term to
long-term),
·
the frequency of interaction (ranging from low to high),
·
the pattern of interaction (ranging from completely
unstructured to strictly hierarchical: peer-to-peer, broadcast, etc.), and
·
the variability of interaction (ranging from fixed to
changeable) as some learning requires only minimal interaction, some maximal.
Involvement-specific features.
(a) The relevance of involvement and (b) role played during
involvement. With respect to relevance,
there are two extremes: the involvement
of an agent is not a conditioned for goal attainment because its learning
activities could be executed by another available agent as well; and the
learning goal could not be achieved without the involvement of exactly this
agent. With respect to the role, an
agent may act as a “generalist” in so far as it performs all learning
activities (like centralized learning, but centralized learning precludes
interaction!) or it may act as a “specialist” – learning a particular activity.
Goal-specific features. (a) The type of
involvement that is tried to be achieved by learning and (b) the compatibility
of the learning goals pursued by the agents. The first feature leads to the important
distinction between learning that aims at an improvement with respect to a
single agent (e.g., its motor skills or inference abilities) and learning that
aims at an improvement with respect to several agents acting as a group (e.g.,
their communication and negotiation abilities or their degree of coordination
and coherence). The second feature leads
to the important distinction between conflicting and complementary learning goals.
The learning method.
·
rote learning (i.e., direct implantation of knowledge and skills
without requiring further inference or transformation from the learner, like
primary/elementary school)
·
learning from instruction and by advice taking (i.e.,
operationalization—transformation into an internal representation and
integration with prior knowledge and skills—of new information like an
instruction or advice that is not directly executable by the learner)
·
learning from examples and by practice (i.e., extraction and
refinement of knowledge and skills like a general concept or a standardized
pattern of motion from positive and negative examples or from practical
experience)
·
learning by analogy (i.e., solution-preserving information of
knowledge and skills from a solved to a similar but unsolved problem)
·
learning by discovery (i.e., discovering and gathering new
knowledge and skills by making observations)
The learning feedback. The learning
feedback indicates the performance level achieved so far.
·
supervised learning (i.e., the feedback specifies the desired
activity of the learner and the
objective is to match this desired action as closely as possible),
·
reinforcement learning (i.e., the feedback only specifies the utility
of the actual activity of the learner and the objective is to maximize this
utility),
·
unsupervised learning (i.e., no explicit feedback is provided and
the objective is to find out useful and desired activities on the basis of
trial-and-error and self-organization processes)
Learning
and communication are related to each other:
(1) Learning to
communicate: Learning is viewed as a
method for reducing the load of communication among individual agents –
communication usually is very slow and expensive, and therefore should be
avoided or at least reduced whenever this is possible.
(2) Communication
as learning: Communication is viewed
as a method for exchanging information that allows agents to continue or refine
their learning activities – learning is inherently limited in its potential
effects by the information that is available to and can be processed by an
agent.
Both
lines of research are related to the following issues:
(1) What to
communicate (e.g., what information is of interest to the others)
(2) When to
communicate (e.g., what efforts should an agent investigate in solving a
problem before asking others for support)
(3) With whom to
communicate (e.g., what agent is interested in this information, what agent
should be asked for support)
(4) How to
communicate (e.g., at what level should the agents communicate, what language
and protocol should be used, should the exchange of information occur
directly—point-to-point and broadcast—or via a blackboard mechanism)
Reducing Communication by Learning
Broadcasting
is costly. Direct communication paths
are not always known.
The
primary idea underlying addressee learning is to reduce the
communication efforts for tasks announcement by enabling the individual agents
to acquire and refine knowledge about the other agents’ task solving
abilities. With the help of the acquired
knowledge, tasks can be assigned more directly without the need of broadcasting
their announcements to all agents.
The
specification of a task is of the form
,
where
is an attribute of and is the attribute’s
value. For each two attributes and the distance between
them is defined as
.
In
the most simplest form, they are defined as
Then
the similarity of the two tasks is:
.
For
every task, , a set of similar tasks, , can be defined by specifying the demands on the similarity
between tasks. An example of such a
specification is
.
Now
consider the situation in which an agent has to decide about assigning some
task to another agent. Instead of broadcasting the announcement of , the agent tries to pre-select one or several agents which
it considers as appropriate for solving by calculating for
each neighbor M the suitability:
where
is an experience-based
measure indicating how good or bad has been performed by M
in the past.
Improving Learning by
Communication
Agents
cannot be assumed to be omniscient without violating realistic
assumptions. In general, agents have
incomplete information about:
(1) the
environment in which it is embedded and the problem to be solved
(2) other agents
(3) the
dependencies among different activities and the effects of one own’s and other
agents’ activities on the environment and on potential future activities.
Two
forms of improving learning by communication:
(1) Learning based on low-level communication, that is,
relatively simple query-and-answer interactions for the purpose of exchanging
missing pieces of information (knowledge and belief) – shared information, and
(2) Learning based on high-level communication, that is,
more complex communicative interactions like negotiation and mutual explanation
for the purpose of combining and synthesizing pieces of information – shared
understanding