By Talia Beechick
It’s an established fact of education that every student learns differently. Some excel with the basic lecture, while others need hands-on practice and still others need to read and think about the topic on their own. It’s also an established fact that catering to the many different learning styles in a classroom of 30 is nigh impossible for teachers, and individual tutoring is too costly for most parents and school systems. But the developers of AutoTutor and Affective AutoTutor have created interactive computer programs that focus on tailoring a child’s education to fit their learning style.
Created by Assistant Professor of Psychology at University of Notre Dame Sidney D’Mello, Art Graesser from the University of Memphis and a fellow researcher at MIT, AutoTutor and Affective AutoTutor respond to students cognitive and emotional needs. Classified as an Intelligent Tutoring System (ITS), the AutoTutor programs aid students in lessons concerning Newtonian physics, critical thinking and computer literacy through real conversations.
AutoTutor plays on the ideal state developed by psychologists to promote cognitive learning and growth. This ideal state features enough stimulation to engage and hold the pupil’s interest while not proving too challenging to become overwhelming for him. Affective AutoTutor adds an emotional component, measuring the emotional state of the student and adjusting its manner of teaching accordingly. Therefore, the program will constantly play a balancing game to either increase the difficulty and eliminate boredom within the pupil, or decrease the level of difficulty to encourage him. It monitors their facial features, body language, posture and conversational cues in order to regulate their mood, and is able to recognize emotions such as frustration, delight, surprise, confusion, boredom and flow (engagement).
Affective AutoTutor uses this information to adjust verbal responses, speech intonation, and facial expressions (of the animated tutor) to respond accordingly. For example, if the program senses the student is frustrated through an upset facial expression or verbal complaint, the animated tutor will change its voice to a more calming tone and perhaps even give a hint to encourage the student to make progress. If Affective AutoTutor senses the pupil is bored by detecting a more slouched posture or even a yawn, it will ask more challenging questions or prompt for more specific information from the student in his response.
These digital tutors could not only revolutionize the way we are able to learn and gain information, but also the way we communicate and interact with computer software in general. This software was inspired by the considerable evidence showing that both human tutors and individualized learning produce better results, and aims to make privatized learning – currently prohibitively expensive for most families – available to all students.
Both programs operate similarly: the process begins with the animated tutor asking the student a complex question and having him write a paragraph-long response. By prompting the student through asking further questions (such as, “What else?”), Autotutor then analyzes his answer to determine his level of knowledge and address any misconceptions that he might have concerning the material. This program also features the ability to respond to questions posed by the student, and maintains the student’s interest through the incorporation of images, animations and simulations into the lesson.
To further develop this software and continue its progress, the National Science Foundation granted the University of Memphis’ Institute for Intelligent Systems $1.1 million dollars this past August, bumping their total up to about $15 million in research grants.
Since AutoTutor’s inception, several systems have been developed which branch off from the original AutoTutor’s goal and subject matter (physics and computer literacy). These systems include GuruTutor, which specializes in biology, HURA Advisor, which emphasizes research ethics, and Operation ARIES, which takes a deeper look at critical thinking in the world of science.
Beginning this year, Operation ARIES will be commercialized through Pearson Education and is one of the first AutoTutor products to make it out of the lab and into the hands of the public. Andrew Olney, associate director of cognitive psychology at the Institute for Intelligent Systems, claims that this program’s lack of availability to the public is due to technical challenges; researchers need a company to license the product in order to create the specific lessons desired by the customer (i.e. a principal at a school).
Other programs, such as AutoTutor-Lite, encompass a broad range of subjects to assist students with. This program in specific is being licensed to corporations and universities, and has been continually updated to ensure the most accurate and relevant information. These systems also differ in terms of executing their lesson plan; while some, such as the original AutoTutor, focus on dialogue between the animated tutor and pupil as the key to successful instruction. Others, such as Operation ARIES, feature a “trialog” method in which the human student, animated tutor and animated student agent participate in educational games designed for high school and college students.
These different teaching methods aim to target different types of learners; researchers recognize that some students prefer an one-on-one, individualized session with a watchful tutor to guide them through the lesson, whereas others favor a more relaxed, casual atmosphere with educational games and other students to learn alongside.
The positive results from AutoTutor and Affective AutoTutor can be attributed to the fact that they model the dialogue patterns, conversational language, bodily gestures and pedagogical methods of top-notch human tutors. AutoTutor has been tested on over 1,000 students and has already seen great success; it outperformed most human tutors and, on average, produced improvements of an entire letter grade in students. This makes it comparable to expert human tutors, proving our advancements in the world of computer intelligence and how it can be used to further our own intelligence, one pupil at a time.