INITIAL MEETING 1:00 - 2:20 THURSDAY, MAY 1,
2003
in HC 25 (the ADI room)
http://www.cs.unb.ca/~bspencer/cs6905ateb/proposal.htm
IF INTERESTED COME ON THURSDAY AND/OR EMAIL BSPENCER@UNB.CA
The web is increasingly seen as both a facilitator of e-business communication and a repository of knowledge about business transactions. Between these two roles a wide gap still exists. People are starting to look to the web to facilitate their e-business transactions, not just to communicate with business partners, but also to use that reposited knowledge to guide their transactions. This requires using the knowledge for tasks not originally intended: interpreting it, acting on it, making decisions from it, using judgement, and recovering from errors and miscommunication. It also means inferring new knowledge from the large amount of business data that is and will be available.
In this course, we study three concrete technologies that address this overall need. In the first segment, reasoning systems for the e-business are considered which directly apply rule-based knowledge to information directly observed. In the second, qualitative and diagnostic reasoning are studied, comparing the behaviour of an actual system to that of an ideal model described at a higher level; this can apply when e-business systems do not behave as expected and faults need to be identified. In the third, business intelligence and high speed data processing is introduced for creating knowledge from data, so that past experience can support decisions affecting future e-business transactions.
The topics for each segment are below. Each segment will contribute one third of the course grade and will involve one or two assignments and a test. For each segment one third of the mark will depend on assignments and two thirds on the test. Assignment dates will be announced. The test dates are shown below.
The schedule will be announced. An organizational meeting will be held 1:00 – 2:30 PM in the ADI Room, Thursday May 1. The schedule will be fixed at that meeting. Segments 1 and 2 will meet twice per week and segment 3 will meet one day per week.
Offered by Bruce Spencer, NRC Group Leader and FCS Adjunct Professor, bspencer@unb.ca or Bruce.Spencer@nrc.gc.ca phone 506-451-2559.
This course will focus on the main issues addressed by the Internet Logic group at IIT e-business: the study of formal languages for expressing knowledge and intentions about business situations on the web, and the creation of systems that draw conclusions and take actions from such knowledge and intentions. Because we use formal languages, we have explicit justifications for our decisions and actions. These are required for producing confidence in e-business, but missing in current e-business systems.
We will briefly cover knowledge representation in the form of first order logic, description logic and ontologies, which form the logic basis of W3C's current Semantic Web proposal. We will rely on RuleML, a proposed W3C and Oasis (www.openoasis.org) standard. The central material of the course will be the design, usage and future development of reasoning systems that are "web-ized" for e-business. Use cases covering policies for privacy and e-commerce will be made more concrete through the several Java Deductive Reasoning Engines for the Web (jDREWs), namely the top-down, the bottom-up, the reactive and the inference queue deductive reasoning engines, and if time allows, the inference queue and object oriented reasoning engines.
|
Lecture |
Date |
Material |
References |
|
1 |
May 1 |
Course Overview. Internet Logic Group. |
|
|
2 |
May 6 |
Clausal Logic 1: Definite clauses, syntax, semantics, “web-ized” logic, RuleML. |
|
|
3 |
May 8 |
Clausal Logic 2: Logical consequences, inference rules, soundness and completeness. |
Online Notes |
|
4 |
May 13 |
Clausal Logic 3: Efficiency vs. expressiveness, description logic. Ontologies and the Semantic Web. |
Online Notes, dl.kr.org/ www.w3.org/2001/sw/. |
|
5 |
May 15 |
Case study: Eligibility rules for e-commerce, jDREW – TD top down, goal trees, OASIS and Policy RuleML. |
www.oasis-open.org, clip.dia.fi.upm.es/COLOGNET-WS/2002/PAPERS/BruceSpencer.pdf. |
|
6 |
May 20 |
Case study: P3P and APPEL, privacy rules for e-health, jDREW – BU, bottom up. |
Online Notes. www.w3.org/p3p, www.w3.org/TR/P3P-preferences/ |
|
7 |
May 22 |
McCune’s discrimination trees and jDREW. Indexing the web, RDF. |
Online Notes, www.w3.org/RDF. |
|
8 |
May 27 |
jDREW – IQ, inference queue; jDREW – OO object oriented. Course Summary. |
Online Notes. |
|
9 |
May 29 |
Test |
|
There will be two assignments and one test.
Additional References:
Lecture 2&3: www.cs.bris.ac.uk/%7Eflach/SL/slides/SLchapter2HTML/index.htm
Lecture 5: Keping Jia, and Bruce Spencer, Negotiating Exchanges of Private Information for Web to appear AI'2003: The Sixteenth Canadian Conference on Artificial Intelligence, Halifax, June 11, 2003.
Lecture 7: William McCune, Experiments with Discrimination-Tree Indexing and Path Indexing for Term Retrieval, Journal of Automated Reasoning, Vol. 9, no. 2 (1992) 147-167. On Reserve.
Model-based Reasoning and Qualitative Reasoning
Offered by Yuhong Yan, NRC Research Officer
and FCS Honorary Research Associate Yuhong.Yan@nrc.gc.ca,
phone 506-452-3975 ,
and Philippe Dague, université Paris Nord (email and phone TBA)..
Model-based Reasoning is an inferring process using models abstracted from the reality of a system. MBR is the symbolic processing of an explicit representation of the internal working of a system in order to predict, simulate and explain the resultant behaviour of the system from the structure, causality, functional and behaviour of its components.
Starting from 1980s, MBR is now successfully used to design, analyse, simulate, diagnose, monitor and maintain a physical technical system. These techniques can significantly improve the efficiency of solving engineering problems. These techniques can also be applied to other application domains such as education, e-business, ecology, biology, and medicine.
The theoretical prologue of this session is first-order logic and reasoning techniques (will be covered in the first part of this course). Then we will focus on the techniques like assumption-based truth maintenance, constraint propagation, qualitative reasoning. Use cases will be an engineering system and an e-business system (one for company performance analysis, another from the New Brunswick Opportunities Network project under investigation).
|
Lecture |
Date |
Content |
|
1 |
June 3 |
Course Overview, Theory of Diagnosis from First Principles |
|
2 |
June 5 |
Assumption-based TMS and diagnosis engine |
|
3 |
June 10 |
Reasoning in presence of fault modes |
|
4 |
June 12 |
Non minimal Diagnosis |
|
5 |
June 17 |
Non minimal Diagnosis |
|
6 |
June 19 |
Fundamentals of Qualitative Reasoning (sings, absolute orders of magnitude, relative orders of magnitude, interval arithmetic) |
|
7 |
June 24 |
Fundamentals of Qualitative Reasoning (QSIM) |
|
8 |
June 26 |
Use Cases in e-business |
|
9 |
July 3 |
Test |
There will be one test, which may be an oral presentation, depending on the number of students.
References:
Walter Hamscher, Johan de Kleer, and Luca Console (eds.) Readings in Model-based Diagnosis, Morgan Kaufmann, 1992.
Feelders and Daniels, A general model for automated business diagnosis, European Journal of Operational Research, Elsevier, 2001, p623-637.
Feelders A.J. and Daniels H.A.M., A formal framework for diagnosis in business performance, Proceedings of the International Conference on Intelligent Systems, pp.123-134, Singapore, (1994).
Offered by Daniel Lemire, NRC Researcher
and UNB Math Adjunct Professor, Daniel.Lemire@nrc.gc.ca,
phone 506-451-6081, and
Owen Kaser, Department of Computer Science & Applied Statistics at UNB
Saint John, owen@unbsj.ca, phone
506-648-5728.
The schedule for this segment may change to two lectures once per week.
On-line Analytical Processing (OLAP) aims to accelerate typical queries in large databases or data warehouses so that on-line performance is possible. OLAP constitutes one of the core data-mining technologies used in industry. This section of the course will target the core technology behind OLAP.
|
Lecture |
Date |
Material |
|
1 |
July 8 |
OLAP : relational databases, motivation, examples, attributes, indexes, on-line queries and the data cube, overview of commercial products |
|
2 |
July 10 |
Data models including Multidimensional OLAP (MOLAP), Relational OLAP (ROLAP), and Hybrid OLAP (HOLAP), star and snowflakes schemas |
|
3 |
July 15 |
Basic OLAP operations including granularity (Roll-up, Drill-down), Slice and Dice, and Pivot. Their complexity analysis. Comparison with relational databases. |
|
4 |
July 17 |
Data Mining using OLAP: deductive analysis using OLAP, testing relationships, association rules (support and confidence), attribute-oriented mining, how data mining using OLAP differs from inductive data mining, statistics vs brute-force OLAP, iceberg queries. |
|
5 |
July 22 |
Implementation issues : memory management, populating data cubes, implementing ROLAP and MOLAP |
|
6 |
July 24 |
Advanced topics (Owen Kaser): Parallel processing: Parallelism in both cube construction and data mining |
|
7 |
July 29 |
Advanced topics (Daniel Lemire): Range queries: polynomial range sum queries and wavelets. |
|
8 |
July 31 |
Test |
Alex Berson and Stephen J. Smith, "Data Warehousing, Data Mining, and OLAP",
McGraw Hill, 1997.
J. Gray, A. Bosworth, A. Layman, H. Pirahesh. Data cube: A relational aggregation operator generalizing group-by, cross-tabs and subtotals. In Proc. ICDE 1996, pages 131-139, 1996.
Erik Thomsen, OLAP Solutions: Building Multidimensional Information Systems, Second Edition, John Wiley & Sons, 2002, 661 pages (ISBN 0471400300). (See http://library.books24x7.com/book/id_3651/toc.asp for TOC and full content.)
There will be one programming assignment (TBD) and one final test.