Presentation on theme: "AH Example Systems Dr. Alexandra Cristea"— Presentation transcript:
AH Example Systems Dr. Alexandra Cristea firstname.lastname@example.org http://www.dcs.warwick.ac.uk/~acristea/
Example Adaptive Hypermedia Systems We show examples that are very different: –TV Scout: personalized TV guide (GMD Darmstadt) –AIMS: Adaptive Information Management System (TU/e+UT) –SQL-Tutor: Intelligent Tutoring System for SQL (Canterbury, –New Zealand) –ISIS tutor (Moscow State University) –Interbook: Adaptive Electronic Textbooks (Univ. of Pittsburgh) –INTRIGUE: adaptable tourist guide (Univ. of Torino) –HERA –MOT (My Online Teacher) –Adaptation to learning styles in (an extension of) AHA! –The ARIA Photo Agent (MIT) with commonsense reasoning
TV Scout: Personalized TV Guide A cooperation between GMD-IPSI and –Goal: Help users in creating their personal TV schedule –Short-lived data (not a static database) –Low user effort required to “tune” the system –Filtering based on time and genre, information provided by the stations –Users plan only for one day –TV Scout has a simple and an advanced interface, with possibilities for collaborative filtering.
suggest queries program description list program description table viewing time profile editor channel profile editor video labels laundry list QSA profile editor QSA profile editor (experts) text search query menus QSA menu TV Scout user interface with starting page retention menus
TV Scout: Evaluation / Feedback Orientation is easy, but undo is missing For some users the system is still too complex (opening folders, buttons to small for visually impaired users) People liked the „grocery list“ (forms interface) Overall it is useful and easy to use High fun-factor! Biggest success indicator is repeat visits by users
AIMS: Task-Based Information Retrieval Agent-based Information Management System: –concept visualization (using “aquabrowser”) –task-based search (keyword search extended with task information) –user model: keeps track of user’s knowledge and performed tasks –graphical user-interfaces for creating concepts, tasks, courses, etc. –initiated at and evaluated with students from the Universiteit Twente Note: adaptation to a “moving target”, because the knowledge changes
AIMS Global Information Model Domain model: defines subject domain by means of a concept map –concepts are linked to each other (“ontology”) Library model: defines relationship between documents and concepts –how relevant is a document for a given concept Course model: course topics and tasks –tasks are described using concepts, task description, prerequisites, task status Learner model: what the user has learned: –course tasks, domain concepts, library documents –overlay model –built jointly by the user and the system
SQL Tutor Knowledge-based tutor for the SQL language – based on constraint-based modeling – currently deals only with the SELECT statement – users register with an initial knowledge level – system suggests problems based on the knowledge level (based on which clause select, from, where, group by, having or order by the user needs to practice – system was evaluated to find out whether it was useful and pleasant to use – SQL-Tutor is described (and sometimes accessible) at: –http://www.cosc.canterbury.ac.nz/tanja.mitrovic/sql-tutor.htmlhttp://www.cosc.canterbury.ac.nz/tanja.mitrovic/sql-tutor.html –Try out: http://ictg.cosc.canterbury.ac.nz:8000/sql-tutor/login http://ictg.cosc.canterbury.ac.nz:8000/sql-tutor/login
ISIS-Tutor: adaptive annotation/ hiding Tutor for CDS/ ISIS library system –CDS/ISIS is a library system for PCs sponsored by UNESCO –ISIS Tutor developed by Peter Brusilovsky and Leonid Pesin –Descendent from an older system ITEM/P (Moscow State Univ.) –Domain and student model for monitoring student knowledge –Tutor component to perform adaptive task sequencing –Hypertext component lets students navigate through course material. –Learning environment lets users interact with ISIS –Versions to determine learning effect of using adaptation –http://www.cs.joensuu.fi/~mtuki/www_clce.270296/Brusilov.h tmlhttp://www.cs.joensuu.fi/~mtuki/www_clce.270296/Brusilov.h tml
ISIS Tutor with Link Annotation The wrong example:
Interbook tool for adaptive electronic textbooks (developed mostly at the Carnegie Mellon University): –authoring through Microsoft Word (+conversion tools) –domain model: concepts and prerequisite relationships –user model: overlay model, updated through “outcome concepts” of read pages –adaptive link annotation –several additional tools: index, glossary, “teach me” –a good description of Interbook: –http://ausweb.scu.edu.au/aw97/papers/eklund/paper.htmhttp://ausweb.scu.edu.au/aw97/papers/eklund/paper.htm
Interbook: Evaluation Goal: to find a value of adaptive annotation –Electronic textbook about ClarisWorks –25 undergraduate teacher education students –2 groups: with/without adaptive annotation –Format: exploring + testing knowledge –Full action protocol Results: –Sequential navigation dominates (“continue” button) –Adaptive link annotation encourages non-sequential navigation –Most students follow the “green” links
Intrigue: adaptive tourist guide Allows for the planning of a trip –stereotype user modeling –allows to plan a trip for a diverse group, for instance parents with children –takes physical disabilities into account, age, interests, etc. –can produce output in html or wml (for mobile phone) –can sometimes be tried at: –http://silk.di.unito.it:8083/ishtar/intrigue.htmlhttp://silk.di.unito.it:8083/ishtar/intrigue.html
MOT (My Online Teacher) Authoring environment based on the LAOS authoring framework that specifies separation of concerns
MOT (old): Domain Concept attribute creation Current concept conceptattribute Try at: http://e-learning.dsp.pub.ro/mot/http://e-learning.dsp.pub.ro/mot/
Orderingoflessons Weights of sublesson Labels of sublesson MOT (old): Editing a Goal Map
Evaluation of early MOT (2004) Goal point of view evaluationCollaboration - Problems? - Suggestions for solving? How did you do it? - Good points? Completeness (LAOS two layer) - Perceived percentage? expressivity? - (perceived) connectivity degree? Should there be more connections, or less? What extra connections? What superfluous? Adaptivity - How much adaptivity to the design goal is perceived? Design range - How much more can be achieved compared to linear model?
USI point of view evaluation Ease of use - information display, information order, - distance of search (depth); - color scheme, ease of access, ease of installation Robustness - parallelism (data overlap), security, recovery Complexity - analysis of possible reduction.
Semantics in MOT MOT is based on LAOS and on semantic web directives necessary –more explicit ontologies, –synonyms - to identify semantic overlaps
User Model in MOT dynamic model of the user's history; user model variables should be also user writable (flag); retrieved by prompting the user –Specified where? AM? UM?
Evaluation of new MOT (2007) intensive two-week course AH & SW 33 out of 61 students selected: 4 th year Engineering & 2 nd year MsC in CS theoretical exam half way for selecting students due to space constraints in computer room at the end: practical exam & 5 questionaires 3 systems: OLD MOT, NEW MOT & Sesame2MOT
Questionnaires SUS questionnaire for comparing usability Multiple choice questionnaire constructed direct questions based upon division of main hypotheses
Second-Order Adaptation Most systems adapt to one parameter: –Recommender systems adapt to what they think the user’s interests are –Learning systems adapt to what they think the user’s knowledge is about certain concepts –Some systems can perform adaptation to devices or network performance More advanced systems adapt to more than one parameter at once –We look at the adaptation to learning styles in an already adaptive learning application
LAG-XLS: an XML Learning Style Adaptation Language Based on the LAG language Elements of the language: –select – selecting concept representation –sort – sequencing concept representation –setDefault – setting defaults –action – updating the User Model
Information about Verbalizer/Vizualizer(Imager) LS
The ARIA Photo Agent (video) Adaptive Linking between Text and Photos –Text is used for searching as it is typed –Text is matched with photo descriptions keywords, people, place and time –Database with “common sense” used –Adaptive sorting (of photos = search results) –Automatic annotation of selected photos –Annotation (conceptual descriptions) of photos can be manually updated –Project webpage: –http://web.media.mit.edu/~lieber/Lieberary/Aria/Aria-Intro.htmlhttp://web.media.mit.edu/~lieber/Lieberary/Aria/Aria-Intro.html
ARIA Screenshot Video at: http://web.media.mit.edu/~lieber/Lieberary/Aria/Commonsense-Aria-Demo.swf http://web.media.mit.edu/~lieber/Lieberary/Aria/Commonsense-Aria-Demo.swf