|
|
|
|
|
|
New to DAML |
|
Reviewed current proposals for temporal
ontologies |
|
Prior work on temporal reasoning: |
|
Allen (1983): Interval Temporal Logic |
|
Allen and Pelavin (1986): Logic of Plans in
Temporally Rich Domains |
|
Allen and Koomen (1989): Planning with a
Temporal World Model |
|
Allen, Kautz, Pelavin, and Tenenberg (1991):
Reasoning about Plans |
|
Ferguson (1994): KR&R for Mixed-Initiative
Planning |
|
Allen and Ferguson (1994): Reasoning about
Actions and Events |
|
Allen, et al. (1993-1997): TRAINS dialogue-based
planning systems |
|
Allen, et al. (1998-present): TRIPS: The
Rochester Interactive Planning System |
|
|
|
|
|
|
Develop practical temporal ontology in DAML |
|
Abstracted from natural language |
|
Dates: “January 3, 2002,” “next Wednesday,” … |
|
Durations: “drive for twenty minutes,” “wait
about an hour,” … |
|
Ordering: “the foundation must be completed
before the walls are started,” … |
|
Constraints: “don’t touch the button while the
switch is on,” “only one flight can use the runway at a time,” ... |
|
Based on Interval Temporal Logic (Allen, 1983) |
|
Solid theoretical foundation |
|
Extensively investigated in the KR literature |
|
Applied to real-world planning and scheduling
problems and natural language applications |
|
|
|
|
|
Approach: |
|
Reify ITL relations into a set of categories
that capture ordering, metric and date information with a simple, intuitive
set of primitives |
|
Example: “Airplane A arrived one hour before
airplane B.” |
|
Deliverables |
|
Practical temporal ontology in DAML |
|
Including “User’s Guide” illustrating use of the
ontology |
|
|
|
|
|
|
Subjective: |
|
Usability: Whether people can use the ontology,
how comfortable they are using it, etc. |
|
Measure via end-user questionaires and
interviews |
|
Although subjective, this is still an important
measure |
|
Objective: |
|
Coverage: What proportion of the end-users’ data
can be represented using the ontology? |
|
Reliability: How likely are different end-users
of the ontology to to encode the same information using the same terms? |
|
For both of these, need multiple end-users
encoding the same information to allow comparison |
|
Inter-encoder reliability seems particularly
important |
|
|
|
|
|
|
Most significant accomplishments to date (5
minutes) |
|
Planned activity in 2002 (5 minutes) |
|
What we plan to deliver in 2002 (technical not
administrative – 3 minutes) |
|
Our metrics for measuring progress (2 minutes) |
|
Time for questions while next speaker is setting
up (3 minutes) |
|
Notes