ITS-MathPS

PGC2018-096463-B-I00

A major objective of this cross-disciplinary proposal is to develop tools, theories and methodologies aimed at improving the teaching and learning of arithmetic-algebraic word problem solving. To this end, we will make use of an existing Intelligent Tutoring System (ITS) that was developed as a result of a joint effort between personnel of the Departments of Computer Science and Didactics of Mathematics of the University of Valencia. This system was the first one to seamlessly consider multiple solution paths without sacrificing guidance capacity, and yet supported both the arithmetical and algebraic way of solving word  problems.

Beyond the possibilities of the system as a teaching tool, and from a  research point of view, this ITS allows us to gather data that helps understand how students learn in the particular field of mathematical word problem solving. In between other aims, this data will be analysed by using machine learning and statistical methods to a) detect patterns of common difficulties, mistakes and misconceptions; b) determine the effect of key instructional actions, e.g. the selection of the proposed task or the level of help provided, not only in a cognitive way but also on other affect-related variables, e.g. motivation; c) discover cultural, social, gender, cognitive and/or affective features that influence learning; d) create models that allow users to be characterized and grouped in categories that are relevant from the point of view of organizing word problem learning trajectories; and e) identify learners with special educational needs, along with appropriate interventions that help them learn better. In addition, our research will not be exclusively focused on a quantitative methodology. The inference mechanism that supports the operation of the ITS is able to check the validity of each solution step, and provides a  privileged scenario to observe the solvers decisions when they are facing difficulties. In these situations, interpretative analysis will allow us to shed light on the possible cognitive processes that triggered the learners mistakes. The results of this analysis will be used to contribute to key methodological aspects, in the form of hypothesis, explanatory models and proposals related to how the instruction should proceed to enhance learning e.g. sequencing of activities, provision of assistance. Both the in class and the one to one case will be studied, taking into account the potential adaptation of the instruction according to the learner characteristics in the one to one case. Hypothesis, theories and proposals will be initially tested by using the ITS and confirmed in real settings when the student resorts to paper and  pencil.

The results generated in this project will help determine best practices in the teaching of word problem solving. In addition, and in more practical terms, we humbly expect that we contribute to shaping the new generation of learning systems, that will surely be aimed at enhancing both learning and users satisfaction by adapting to the student particularities, including people with special needs.

Intelligent Tutoring System, ITS, Problem Solving, Algebra, Arithmetics, Supervision, Scaffolding

José Antonio González Calero

Carlos Soneira Calvo

Cristina Cunha Pérez

Aladdin Ayesh 

Naeem Ramzan 

Pablo Arnau González 

Miguel Ángel García Moreno 

Guillermo Chicote Huete 

Aleix Peiró Rodríguez

Javier del Olmo Muñoz

Sergio Tirado Olivares 

Gema Pérez Buj 

Miguel Arevalillo Herráez

Departament d'Informàtica

Escola Tècnica Superior d'Enginyeria

Universitat de València

Avda. de la Universitat s/n

46100. Burjassot. Valencia

Grant funded by MCIN/AEI/ 10.13039/501100011033 and by “ERDF A way of making Europe”.

Universidade da Coruña

Universidad de Castilla-La Mancha

• AGE - Knowledge Generation Projects