Tutorial on
Context Awarness and Knowledge Traceability
Instructor
|
Nada Matta
University of Technology of Troyes
France
|
|
|
|
Brief Bio
Nada Matta is Professor at the University of Technology of Troyes. Studies techniques in knowledge engineering and management and specially to handle cooperative activities. She assumed several reponsibilities (director of Human, Environment and Technology of Information and Communication Department, Director of Scientific group of Safety and Security, Director of Information Systems Department). She organized several workshops and did Tutorials in Knowledge management jointly to KMIS, IJCAI, ECAI, CTS, CSCW conferences. Nada Matta did her PhD in knowledge engineering and Artificial Intelligence at University of Paul Sabatier in collaboration with ARTEMIS. Worked for four years at INRIA in projects with Dassault-Aviation and Airbus Industry.
|
Abstract
Context awareness is then proposed in the literature as the ability of the system to detect or examine its surrounding environment and respond accordingly [Liu et al, 2011, Rosenberger et al, 2018. Two main categories of context-awareness can be identified: the active and the passive. In this tutorial, we put on the importance of the context awareness to keep track knowledge from daily activities. We present context awareness and traceability principles and explore several approaches in these domains.
Keywords
Context awarness, Knowledge Management, traceability.
Aims and Learning Objectives
We show the importance to consider Context to put on the semantic representation of knowledge produced in daily activities.
Target Audience
Researchers and Industrials.
Prerequisite Knowledge of Audience
None.
Detailed Outline
- Context: Principles and components:
- Traceability: principles;
- Knowledge traceability approaches.
Tutorial on
Multilayer Networks for Modeling and Analysis of Complex Data
Instructor
|
Sharma Chakravarthy
University of Texas at Arlington
United States
|
|
|
|
Brief Bio
Prof. Sharma Chakravarthy is an ACM Distinguished Scientist and ACM Distinguished speaker. He is also an IEEE Senior Member. He is a Fulbright specialist. He has spent several summers at the Rome Air Force Research Laboratory (AFRL) as a Faculty Fellow working in continuous query processing over fault-tolerant networks and video stream analysis. Sharma Chakravarthy is Professor of Computer and Engineering Department at The University of Texas at Arlington, Texas. He established the Information Technology Laboratory at UT Arlington in Jan 2000 and currently heads it.
His current research includes big data analysis using multi-layered networks, stream data processing for disparate domains (e.g., video analysis), scaling graph mining algorithms for analyzing very large social and other networks, active and real-time databases, distributed and heterogeneous databases, query optimization (single, multiple, logic-based, and graph), and multi-media databases.
|
Abstract
Abstract
Over the last two decades, there has been a need to analyze and deduce useful knowledge from a variety of data being collected in every walk of life - social, biological, healthcare, corporate, climate, to name a few. Evidently this complex data consists of multiple entities, features, and relationships giving rise to various modeling, analysis, and visualization challenges. The challenge to model and analyze large complex, disparate data for a broad set of analysis objectives differentiates big data analytics from mining. Hence, flexibility of modeling and analysis is important. Concomitantly, efficiency is important due to large amounts of data and complexity of analyses.
The goal of this tutorial is to provide the reader with an understanding of data modeling and analysis approaches using graphs. Graphs are not new, but how they are used for big data analytics is going through a transformation which is important to understand. The prime focus of this tutorial will be Multilayer Networks (or MLNS) – an area that has gained interest in the recent past due to its ability to model different aspects of complex data sets through the inter-linked set of networks (or simple graphs). Among several ways to analyze MLNs, the novel divide-and-conquer based decoupling approach is flexible, efficient, allows parallelization, and can deal with arbitrarily large data sets (scalability.) This tutorial will introduce MLNs as a modeling alternative and the decoupling approach for analysis, such as community, centrality, and substructure discovery. Using the dashboard developed at UT Arlington, we will do dome hands-on analysis, drill-down, and visualization as well using the publicly available tool at https://itlab.uta.edu/mln-dash/live/
Keywords
Multilayer networks, Modeling, Analysis of various graph metrics, drill-down, Visualization
Aims and Learning Objectives
1. Understand the use of graphs and specifically multilayer networks (MLNs) for modeling complex data.
2. Use of EER (Extended Entity Relationship) approach for modeling MLNs
3. MLN analysis alternatives including the decoupling approach
4. An analysis algorithm (community, substructure, centrality) using the decoupling approach
5. Hands on use of MLN Dashboard (MLN-geeWhiz) to generate layers, analyze, drill-down, and visualize
Target Audience
Masters’ and PhD students engaged in mining and big data analysis research. Academics to understand the scope of MLNs for research and developing new algorithms. Practitioners for applying modeling and analysis using MLNs for their data. Understand/Interpret results using visualization
Prerequisite Knowledge of Audience
Some knowledge of graphs and data modeling and analysis. MS and PhD students and even BS students can benefit from this tutorial. Certainly academics and practitioners. Anyone who is interested in knowing current trends in complex data modeling and analusis!
Detailed Outline
1. What is complex data in terms of entities, features, and relationships? What are analysis objectives? Data Analysis life cycle
2. Introduction to simple and attributed graphs and their limitations for modeling complex data
3. Introduction to MLNs (all types), formal definition, and their advantages and disadvantages
4. Analysis alternatives for MLNs and their discussion
5. The decoupling approach and its advantages and disadvantages
6. Examples of developing MLN analysis algorithms using the decoupling approach
a. Community detection
b. One centrality detection
7. Introduction to MLN-geeWhiz dashboard (needs laptop and Internet).
8. Hands-on use of MLN-geeWhiz for some applications (e.g., Airlines data, UK accident data, or DBLP data)
9. Layer generation, analysis, and visualization using MLN-geeWhiz
10. Conclusions