CSI Topics in Fuzzy Systems : Life Log Management Fall Semester, 2008
Teaching Staffs 4 Professor: Cho, Sung-Bae (C515; ; 4 Web page: 4 Class hours –Tue 2, Thu 3, 4 (A542) 4 Office hours –Tue 5, 6 (C515) 4 TA: Hwang, Keum-Sung ( ;
Uncertainties in Intelligent Systems 4 Dealing with uncertain and imprecise information has been one of the major issues in almost all intelligent system –Decision making systems, diagnostic systems, intelligent agent systems, planning systems, data mining, etc 4 Various approaches to cope with uncertain, imprecise, vague, and even inconsistent information –Bayesian and probabilistic methods, belief networks, softcomputing, etc 4 Softcomputing –Neural networks, fuzzy theory, approximate reasoning, derivative-free optimization methods (GA), etc –Synergy allows SC to incorporate human knowledge effectively, deal with imprecision and uncertainty, and learn to adapt to unknown or changing environments for better performance intelligent systems to mimic human intelligence in thinking, learning, reasoning, etc
Life Log MS SenseCam & MyLifeBitsTokyo University (Aizawa)
Life Blog Nokia LifeblogQueens Univ., eyeBlog
Life Log: Related Works 4 Microsoft Research, MyLifeBits 4 Microsoft Research, MemoryLens (PhotoViewer, LifeBrowser) 4 Microsoft Research, JamBayes 4 Nokia, LifeBlog 4 Helsinki University, ContextPhone 4 Carnegie Mellon University, Context-Aware Phone 4 MIT Ambient Intelligence Group, PhotoWhere 4 MIT Reality Mining Group, Serendipity Service 4 MIT Reality Mining Group, Interactive automatically generated diary 4 …
Difficult to collect large volume of heterogeneo us data Difficult to access and retrieve data Lack of meta data format and visualization for life log Difficult to make high- level tags Life Log Key Issues
Course Objectives 4 Introduce the several techniques on fuzzy systems & pattern recognition, and case studies on recent applications for the management of life log 4 Cover the themes to manage life log –Log collection –Preprocessing –Recognition and inference –Application services 4 Acquire relevant knowledge through course projects
Syllabus 1. 9/2, 4: Course Introduction & Overview of Life Logging 2. 9/9, 11: Sensor Data Collection 3. 9/16, 18: Preprocessing 4. 9/23, 25: Feature Extraction 5. 9/30, 10/2: Classification 6. 10/7, 9: Bayesian Networks 7. 10/14, 16: Term Project Proposal 8. 10/21, 23: Midterm Exam 9. 10/28, 30: Dynamic Bayesian Networks /4, 6: Dynamic Bayesian Networks (SL-1) /11, 13:Hidden Markov Models /18, 20: Ontology and Context Modeling (SL-2) /25, 27: Emotion/Activity Recognition (SL-3) /2, 4: Lifelog Management and Visualization /9, 11: Term Project Final Presentation /16, 18: Final Exam (TBD)
Evaluation Criteria 4 Evaluation Criteria –Term Project (written report & oral presentation): 50% –Written Exam: 30% –Presentation: 20% 4 Term Project (Oral presentation is required) : –Theoretical Issue (analysis, experiment, simulation) : Originality –Interesting Programming (Game, Demo, etc) : Performance –Survey : Completeness