Intelligent Tutors for All the Constraintbased Approach Tanja

Intelligent Tutors for All: the Constraint-based Approach Tanja Mitrovic Intelligent Computer Tutoring Group University of Canterbury

Intelligent Tutoring Systems § Goal: one-to-one teaching without the expense of human tutoring § Simulate a human teacher § Problem-solving environments (learning by doing) § Based on Artificial Intelligence § Student modeling

Student modeling Co ll ec ts Data about user System s se s e c o Pr Ada pts Student Model Adaptation Effect

Architecture of ITSs Domain knowledge Domain module Student modeler Pedagogical module Pedagogical expertise Student Models Interface Communication knowledge Student

Learning from performance errors § Ohlsson, 1992 § Declarative/procedural knowledge § Constraints as a knowledge-representation formalism § Constraints do not assert anything § Constraints encode correctness for a domain § “If the relevance condition R is true, then the satisfaction condition S ought to be true, otherwise something is wrong. ” § Constraints support judgment, not inference

Learning from performance errors § Learning phases: § Error detection § Error correction § How can we catch ourselves making errors? § If the knowledge is there, then why the error? § If not, then how is the error detected? § CBM: domain and student modeling

Constraint-based Modeling § The space of incorrect knowledge is vast § Therefore: abstractions are needed § Represent only basic domain principles § Group the states into equivalence classes according to their pedagogical importance

Constraint-Based Modeling § Domain knowledge represented by a set of constraints § A constraint is a pattern of form <Cr, Cs> § If a solution matches the Cr then it must also match the Cs, else something is wrong § “Innocent until proven guilty” approach

Example constraints § If you are driving in New Zealand, you better be on the left side of the road. § If the current problem is a/b + c/d, and the student’s solution is (a+c)/n, then it had better be the case that n=b=d.

Advantages of CBM § Very efficient computationally § No need for a problem solver § No need for a bug library § Insensitive to the radical strategy variability phenomenon § Neutral with respect to pedagogy

Implications for ITS Design: CBM § Represent the domain in terms of constraints § Model the student in terms of constraints § Pedagogy: § Augment student’s constraint base § When should the ITS take an initiative? § What to instruction to deliver? § Models of meta-cognitive skills § Student’s meta-cognitive skills

CBM: Model the Student § A violated constraint implies incomplete or incorrect knowledge § Short-term student model: § the set of violated constraints § the set of satisfied constraints § No one-to-one mapping between problems and constraints § Long-term student model: § Constraint histories (overlay/probabilistic)

CBM: Pedagogy § Constraint-based tutors function by augmenting the student’s own knowledge base § Choose practice problems that exercise constraints § Interrupt when a constraint is violated § Attach feedback messages to the constraints § Tell the student which constraint he/she just violated and how

History of ICTG § SQL-Tutor § Solaris (1997), Windows (1998), Web (1999) § CAPIT (2000) § KERMIT (2000) § WETAS (2002) § LBITS (2002) § NORMIT (2002) § ERM-Tutor (2003) § COLLECT-UML (2005) § ASPIRE, VIPER § J-LATTE § Thermo-Tutor

CAPIT

LBITS – elementary vocabulary

Copy Paste Pen Get the pen, each time you want to update the group diagram and Leave it as soon as you are done Group Diagram Chat Area Individual Diagram Feedback Area

Current work § ASPIRE, VIPER § Supporting meta-cognitive skills (selfexplanation, self-assessment …) § Affective modeling and pedagogical agents § Supporting multiple teaching strategies § New ITSs