AISEN

Keywords: dynamic assessment, adaptive assessment, artificial intelligence, graded hints, reading processes, executive functions

Demand for online education has grown rapidly in recent decades. According to UNESCO, in May 2020, more than 1,600 million students will be affected by school closures due to the COVID-19 pandemic. One of the main problems of online education is the assessment of learners. Complex assessment techniques and services are required to be available through online platforms, not only to measure learning skills, knowledge or proficiency, but also to evaluate other constructs such as mental disorders, clinical depression, intelligence, job satisfaction, work stress, etc. All of these constructs require data analysis techniques to infer from learners’ performance in assessments. Another problem with most educational platforms is the teaching process itself. These platforms usually offer the same learning sequence to all their students without taking into account their initial knowledge, their needs or their characteristics. Society is currently demanding new learning services that can mimic the tutoring behavior of a human teacher, supporting the learner’s process and addressing his or her special educational needs (SEN). There are an estimated 200 million disabled children in the world. According to UNICEF, disability is one of the most serious barriers to education worldwide. Traditionally, SEN has been approached as a problem to be solved, rather than identifying and changing the factors that lead to learning difficulties. Differentiated learning is a strategy for inclusive education in which students share the same learning goals and learn similar things.

Today, online learning tools could be seen as providers of self-paced educational content without clear pedagogical strategies and without taking into account learners with SEN. However, it is not only online tools that show this lack; the adaptation of the learning process is also a challenge in schools in most countries. Initiatives such as ‘No Child Left Behind’ in the USA put the emphasis on providing equal educational opportunities for all learners. However, many countries do not have sufficient financial and human resources to provide adequate support for these students, and in many cases SEN diagnosis is delayed, slowing down the learning process. In addition, the assessment tools used by psychologists generally consist of a large number of tasks that have to be administered in their entirety to all these students, which in many cases becomes an exhausting procedure for them. Experts have long criticized conventional assessments, which merely classify, rank and quantitatively compare learners with their peers, for failing to provide information about their learning potential and, overall, for failing to identify the conditions or resources that would favor higher performance in solving test content.

The role of technology in supporting students with SEN is widely recognized. According to the European Commission, the integration of technological solutions into learning removes some of these barriers for children who may need more support in terms of teaching or physical means of learning. The combination of digital technology and innovation in education allows teachers to receive and give real-time feedback, helping them to support students with SEN. This synergy will help to reduce the digital divide caused by various factors such as disability or socio-economic status. However, providing effective support for SEN is still a challenge that this project will try to address. In this line, we will develop a web-based platform for the assessment and optimization of executive functions and reading processes, through activities and games designed by a team of research psychologists who are experts in these fields (mainly from the Universities of Seville and Zaragoza). The platform will also include AAI techniques for adaptive assessment, learning and scaffolding, developed by researchers in Artificial Intelligence in Education at the University of Malaga.

The main objective of the project is to develop and empirically validate an intelligent web-based software environment for the assessment and optimization of executive functions and reading processes in students with neurodevelopmental and learning disorders. The technological development objective is to design and develop a set of interactive activities and games that will be dynamically adapted to the student’s progressive level of competence using AI techniques. These interactive activities will include an intelligent mechanism of graduated prompts adapted to optimize executive functions and reading processes. Their application will therefore make it possible to obtain information on both (a) the level of difficulty shown in solving the activities and (b) the type and degree of support required to solve them successfully. The validation objective involves testing and analyzing the extent to which the use of these activities significantly increases the predictive validity (incremental validity) of the executive function and reading performance experienced by the students, with respect to the prediction made on the basis of the results obtained with statically applied criterion tests. This validation objective also involves testing the effectiveness of the graded prompts in optimizing the learning process.

Principal investigators:

• Eduardo Guzmán (Universidad de Málaga)
• Isabel de los Reyes Rodríguez (Universidad de Sevilla)

Researchers:

• Universidad de Málaga
  • María Victoria Belmonte
  • Eva Millán
  • Mónica Trella
  • Ángel J. Rodríguez Mercado
• Universidad de Sevilla
  • David Saldaña
  • Antonio Aguilera
• Universidad de Zaragoza
  • Juan Carlos Bustamante
  • Elena Escolano
  • Eva Vicente
  • Ester Ayllón
  • M. Ángeles Bravo
  • M. Antonio Acero
  • Fernando Martín
  • Estefanía Ortas
• Nieves Moyano (Universidad de Jaén)