CogSci Summaries home |
S. Russell and E. Wefald (1991). Do the right thing : studies in limited rationality (Chapter 2: Metareasoning Architectures), MIT Press
author = "Stuart Russell and Eric Wefald",
title = "Do the right thing : studies in limited rationality",
publisher = "MIT Press",
year = "1991",
series = "Artificial intelligence",
address = "Cambridge, Mass",
Author of the summary: David Furcy, 1999, firstname.lastname@example.org
Cite this paper for:
- SYSTEMS: TEIRESIAS, MRS and SOAR (a brief discussion of each as a
- Intelligence = acquiring and using declarative knowledge and gradually
compiling it into an efficient execution architecture.
- Metalevel knowledge can be derived from domain-level knowledge and
domain-independent knowledge about the internal operations of the system.
- Taxonomy of knowledge types.
- Taxonomy of execution architectures.
- Compilation as a necessary process for increased efficiency.
- Importance of the coexistence of all types of execution architectures
at both the object-level and metareasoning level with compilation processes
to transform one type of knowledge into another.
Different types of knowledge are identified which define four possible
execution architectures (EA's). After describing both taxonomies of knowledge
and EA's, the authors insist on the importance of compilation processes to
enable an agent to make the same decisions faster. The main claim is that
a bounded optimal agent should be designed using all four types of EA's at
both the object-level and metareasoning level. This will enable the agent
to learn as fast as possible in complex domains with varying degrees of
regularity. In conclusion, metalevel knowledge is NOT a separate kind of
knowledge but instead reduces to knowledge about the operation of the
object-level decision procedure.
- Traditional definition: Reasoning about reasoning i.e., reasoning
about entities that are internal to the system, as opposed to
object-level (or OL) reasoning, which
manipulates symbols referring to external entities.
- In this book: There exists an identifiable decision procedure to choose
computation steps (Note that declarative representations are not required).
Metareasoning occurs when the nature and order of
computation steps in the OL decision procedure depends on current events (not
hardwired at design time).
Why? For efficiency reasons (time constraints) and to choose relevant
- Notion of introspective fidelity, when the recommendations
of the metalevel are always followed by the OL reasoning. Most current systems
(except SOAR) do this.
- Heterogeneous versus uniform metalevel architectures.
The latter enable
infinite regress. This problem explains why optimal control of ALL reasoning
is not possible and justifies the need for default, hardwired and/or
An execution architecture (EA) produces a decision using some knowledge.
Main point: Different kinds of knowledge can be acquired from perception.
Thus, there exist several (in fact 4) types of EA's.
Types of knowledge
The uncompiled architecture uses knowledge of types A,B,C plus the
decision-theoretic (DT) principle to make decisions.
- Static knowledge:
- description of some aspect of a world state -- condition(state)
- action results -- condition(result(action,state))
- utility of world states -- utility(state)
- Dynamic knowledge:
- Type A: condition(s) -> condition'(s)
- Type B: condition(s) -> condition'(result(a,s))
- Type C: condition(s) -> utility(s,v)
- Type D: condition(s) -> best(a,s)
- Type E: condition(s) -> utility(result(a,s),v)
- Type F: condition(result(a,s)) -> best(a,s)
Compilation is the process of producing a more efficient implementation
of an input-output mapping by transforming one type of knowledge into another.
The "most compiled" architectures are called tropistic agents, i.e.
reactive agents using production-like knowledge the conditions of which can
be directly matched with sensory inputs.
Examples of homogeneous compilation: A + A -> A or B + B -> B as is done
in EBL and with macro-operators.
Examples of heterogeneous compilation:
Types of EA's
- Type D or condition-action rules can be compiled using type
B and F knowledge (among other ways).
- Type E or action-utility rules can be compiled using type
B and C knowledge.
- Type F or goals can be compiled from type C knowledge and
the decision-theoretic principle.
Since complex conditions take space for storage and requires time for matching,
approximate compilation can be performed (as in Lazy EBL) to trade off
accuracy and efficiency. Approximate productions can fire rapidly (defaults)
and be overridden by more precise ones if time permits.
- Decision-theoretic systems: A + B + C + DT
- Goal-based systems: A + B + F
- Action utility systems: E + DT
- Production systems: D
Existing metalevel architectures
- TEIRESIAS: Rules infer the absolute or relative value of OL rule
applications but the metalevel lacks a model of the OL decision procedure.
- MRS: The metalevel selects both tasks and the methods to accomplish them
and contains a first attempt at modeling the probability and utility of
the outcomes of computational steps at the OL.
- SOAR: The metalevel consists of conducting full-scale simulations of
OL computational steps and of chunking the results into productions. The
simulations causes SOAR to be inadequate for a bounded rational agent.
Summary author's notes:
Back to the Cognitive Science Summaries homepage
Cognitive Science Summaries Webmaster:
Last modified: Mon Apr 5 09:02:17 EDT 1999