1998 AAAI Fall Symposium on Artificial Intelligence and Link Analysis
Schedule
Home
Description
Organizing Committee
Symposium
Dates, times, and places
Why a symposium now?
Who should attend?
Press policy
About AAAI
Orlando Information
Schedule
Registration
Instructions
Cost
Form
Link Analysis
Background
Examples
Artificial Intelligence
Relevant techniques
Software and Data
Existing software
Useful datasets
References
FAQ
Web sites
Bibliography
Contact
Comments
Subscriptions
General Format
The Symposium will consist of four types of activities: surveys to provide participants with a common base of knowledge from which to work, fifteen short papers presenting current technical work, panel discussions to draw on the expertise of participants, and one working session. Specifically:
- Surveys: Three surveys will acquaint participants with a common base
of relevant background material. Most participants will be familiar
with at least one of the surveys; few will be familiar with all.
"Social Network Analysis" will cover basic techniques developed
by social network analysts to characterize linked data. "Visualization"
will cover approaches to visualizing linked data to best understand
relationships and internal structure. "AI Technologies" will cover
a variety of technologies from artificial intelligence that are
relevant to link analysis. Each tutorial will be 40 minutes with
20 minutes for questions and discussion.
- Short papers: Individual participants will present papers describing their
current technical work. The organizing committee will attempt
to divide papers evenly between purely technical papers and papers
describing applications that stretch the limits of current technologies.
Each paper will be brief (15 minutes) with 15 minutes for discussion
and questions after each paper.
- Working sessions: Working sessions will feature structured discussions moderated
by a member of the organizing committee as well as short presentations
from participants. Working session will assess:
- The state of the art in current systems for link analysis – What
are the basic functions of a link analysis system? What are is
the state of the art in these basic functions? Are there important
functions that are currently beyond the capabilities of current
systems?
- Near-term applications of AI technologies – This session will
identify the "low hanging fruit," in other words, fully-developed
AI technologies that could be put to work in less than a year.
How can the current crop of AI techniques facilitate existing
link analysis applications? What new applications might be enabled
by such technologies?
- Characteristics of successful applications of link analysis –
What makes a successful application of link analysis? What are
the fundamental characteristics of link analysis tasks? What major
roadblocks stand between project conception and successful implementation?
- Longer-term research directions – This session will explore how to build a longer-term "technology pipeline," that is, a research community exploring a set of questions that will enable a new generation of link analysis applications. What are the research areas where advances would most aid link analysis applications? What community resources would most aid research (e.g., datasets, code repositories, bibliographies)? What research areas are unique to link analysis and what research areas will receive support for other reasons? How can research findings most easily make their way into developed products?
- The state of the art in current systems for link analysis – What
are the basic functions of a link analysis system? What are is
the state of the art in these basic functions? Are there important
functions that are currently beyond the capabilities of current
systems?
Specific Events
| 9:00 | Welcome, Introductions, Announcements, Symposium Overview David Jensen and Henry Goldberg, Symposium Chairs |
|
| 9:30 | Transnational threat indications and warning: The utility of network
analysis John T. Picarelli, Pacific-Sierra Research Corporation [PDF | PostScript] |
|
| 10:00 | Link analysis of MEDLINE titles as an aid to scientific discovery Don R. Swanson, University of Chicago Neil R. Smalheiser, University of Illinois |
|
| 10:30 | Break | |
| 11:00 | Survey: Visualization Christopher Westphal, Visual Analytics, Inc. |
|
| 12:00 | Lunch | |
| 1:30 | Hydra: A Database system with facilities for link analysis Robert Ayres , Royal Military College of Science, Cranfield University |
|
| 2:00 | Link analysis in the FinCEN AI System Raphael W.H. Wong, Financial Crimes Enforcement Network, U.S. Department of Treasury [PDF | PostScript] |
|
| 2:30 | Survey: Social Network Analysis David Krackhardt, Carnegie Mellon University |
|
| 3:30 | Break | |
| 4:00 | Immersive visualization for link analysis Nancy W. Grady, Daniel R. Tufano, Raymond E. Flanery, Jr., Oak Ridge National Laboratories [PDF] |
|
| 4:30 | Working Session: Current Systems (All participants) |
|
| 5:30 | Break | |
| 6:00 | Reception | |
| 6:30 | Reception ends |
| 9:00 | Phenomenal data mining and link analysis Donal Lyons, University of Dublin Gregory S. Tseytin, V.I. Smirnov Research Institute [PostScript] |
|
| 9:30 | Survey: AI Technologies David Jensen, University of Massachusetts, Amherst |
|
| 10:30 | Break | |
| 11:00 | Working Session: AI Technologies (All participants) |
|
| 12:00 | Lunch | |
| 1:30 | Discovering collaborators by analyzing trails through an information
space David W. Payton, HRL Laboratories [PDF] |
|
| 2:00 | Restructuring Databases by Consolidation and Link Formation Henry Goldberg, National Association of Securities Dealers Regulation [PDF | PostScript] |
|
| 2:30 | Working Session: Applications (All participants) |
|
| 3:30 | Break | |
| 4:00 | Using statistical and relational methods to characterize hyperlink
paths Mark Craven, Carnegie Mellon University [PDF | PostScript] |
|
| 4:30 | Towards a theory for a sociable software architecture Emmett Davis and Bonnie Holte Bennett, University of St. Thomas |
|
| 5:00 | Applying link analysis on automatically extracted information
from KNOW-IT Woojin Paik, Elizabeth Liddy, Eileen Allen, Eric Brown, Andrew Farris, Robert Irwin, Jennifer Liddy, and Ian Niles, TextWise Inc. |
|
| 5:30 | Break | |
| 6:00 | Plenary Session | |
| 8:00 | Plenary Session ends |
| 9:00 | Coreference as the foundations for link analysis over free text
databases Breck Baldwin , Inst. for Research in Cognitive Science, University of Pennsylvania Amit Bagga, Duke University [PDF] |
|
| 9:30 | Automatic information extraction from documents: A tool for intelligence
and law enforcement analysts Richard Lee, Sterling Software [PDF] |
|
| 10:00 | How to introduce artificial intelligence to link analysts Mike Hunter, i2 Limited |
|
| 10:30 | Statistical challenges to inductive inference in linked data David Jensen, University of Massachusetts, Amherst [PDF | PostScript] |
|
| 11:00 | Working Session: Research issues and future directions (all participants) |
|
| 12:30 | Symposium ends |