This project requires you to investigate in detail a topic (normally a
paradigm). You are expected to become familiar with the researchers who proposed
and enhanced it, their most significant papers. If possible, also research the
reception of the approach in various research and application communities, its
champions and detractors.
Surveying a field requires significant creative work. You need to get to a
bottom of a paradigm. You need to find the relevant papers - picking ten papers
from the same author will probably not give you the overview. Sometimes existing
surveys can help you in quickly orienteering in the field.
The minimal hope is that by going through this project you will gain experience
about how to explore a research topic.
The maximal hope is that the project will help you find an interesting research subject
on your own.
- Write-up (about 10 pages in the IEEE double column conference format).
- Introduction: presenting the topic. Who introduced it, when.
Sources and predecessors of the approach.
- Short review of (about) 10 papers on the topic your
- Applications of the approach. Commercial success, if any.
- Compare the approach with its "competitors".
- The reception of the approach in the research and application community.
Your subjective opinion of the approach.
- Optional, but very welcome: did this survey give you ideas for future work?
- Bibliography - all the papers you have reviewed, other relevant papers.
- Presentation (20 minutes - about 10 slides)
- Any major cognitive architecture, other than SOAR.
- Auction protocols.
- Game theoretic approaches for e-commerce agents.
- Mobile agents on the internet (in the Telescript or Aglets sense, not mobile robots)
- Agent communication languages. This topic would require you to investigate the
major speach act based languages (KQML and FIPA ACL) and contrast them to a variety of
agent communication methods. Of special interest would be agents which attempt to
communicate using natural language.
- Agents with human-like emotions. This topic is about "hard" affective computing.
You need to find papers where the agents/programs are programmed such that they
claim that the emotions are really implemented.
- Propose your own. You need to convince me that the subject is extensive
enough - e.g. by providing a list of papers you will review, researchers.