- Reasoning and rationality:
How rational choices guide reasoning, and reasoning shapes rational choices.
Ideal theories separable, but realistic combined theories lacking.
- Limited rationality:
qualitative decision theory, group choice theory.
Limits on time and memory.
- Physical theories of mind:
Physical limits. Characterizing cognitive limits with physical concepts.
- Artificial intelligence: Intelligence = rationality +
knowledge (approximately). Concentration on knowledge, but recent
surge on rationality. Increasing integration with COTS.
- Computer science:
Rationality and knowledge absent or
implicit, with important, growing exceptions. In COTS, ``preference'' means
selection of options. Stupidity of computers a byword.
- AI and DB: Preferential representations and reasoning in
planning, monitoring, revision of plans, beliefs, and preferences.
Utility-guided retrieval methods. Utility models underlying agent
architectures. Guardian agents reflecting and representing
- Algorithms and programming: Specifications including
utility models as well as logical properties. Libraries indexed by
utility models. Utility-based software engineering. Utility-aware
or utility-adaptive algorithms.
- Distributed systems: Agents as rational computational
actors. Collaborative adoption of preferences. Grading rationality
resulting from systemic laws. Negotiation. Group and individual
choices and cognitive properties.
- Security, privacy, and survivability: Disclosure and
transaction disutilities guiding boundary controller construction
and operation. Context-sensitive classification systems. Fungible
system architectures preserving overall utility.
- Communications and networks:
Market allocation of QoS under loading and system degradation.
Anomaly analysis and alerting sensitive to systemic utilities.
- Convenient specification languages for qualitative preferences
and structured utilities.
- Acquisition methods for qualitative preferences and probabilities.
- Automatic and semi-automatic construction/compilation of
quantitative measures from qualitative preferences and background
- Utility-indexed libraries of algorithms, reasoning
methods, and architectures.
- Integrating preferential reasoning and knowledge into programs.
- Quantitative measures of degrees of rationality and
levels of knowledge.
Last modified: Fri Apr 2 11:37:33 EST 1999