
Conference’17, July 2017, Washington, DC, USA Antonio Bucchiarone, Tommaso Martorella, and Diego Colombo
Furthermore, the increasing availability of Open Educational Re-
sources (OER)
6
can help reduce teachers’ material preparation and
lower the accessibility barrier in terms of cost and material kind.
Learning is also crucial in the industry. The continuous techno-
logical disruptures are creating job positions in brand new elds
7
.
Entirely new hard skills are required to t the openings. Moreover,
soft skills are among the most sought-after skills because of their
lower trainability and a slower changing pace
8
. However, the oer
seldom matches the demand. This mismatch led to the need for
upskilling
and
reskilling
. The former means teaching employees
new, advanced and valuable skills to match the prole required for
the next step in their current career path. The latter, instead, targets
employees with a prole similar to the one required. It consists
in teaching them adjacent skills and training them for their new,
dierent job. Both of them are fundamental learning activities that
take place in varied (and often dissimilar) environments.
From a student’s perspective, eective teaching means
1-1 tu-
toring9
. It allows teachers to target specic misunderstandings and
necessities with real-time feedback and explanations relevant to the
student’s experiences. On the other end of the spectrum, the most
teacher-eective approach is the one-to-many lecture, where the
teacher prepares the material upfront for being presented to a wide
audience. One of this approach’s main downsides is encouraging
passive learning. A study by K.R. Koedinger et al. shows that the
"Doer Eect" is a causal association between practice and learn-
ing outcomes and that practicing is six times more eective than
reading [
10
]. Another signicant drawback lies in the motivational
aspect.
Active learning
is more eective in learning outcomes and
motivation than passive learning [
2
]. Despite that, a recent article
on PNAS by L. Deslauriers et al. highlights a negative correlation
between actual learning and the feeling of learning in the students
[
5
]. However,
gamication
and serious games gained consensus
as tools to motivate people to engage in benecial activities, even
if seen as unrewarding or tedious [1,4,6,7].
The design of a learning experience should take into account all
of these dierent factors. On the other hand,
a design framework
should not make assumptions about content type, form, de-
livery, and validation while still removing any obstacle be-
tween the teacher, the student, the environment, and the
learning experience.
Individual coaching is rarely feasible due
to poor scalability, whilst one-to-many general training is scalable
but lacks individualization altogether. That is the gap the Person-
alized and Gamied eTutoring System - PolyGloT - aims to ll.
Adaptive eTutoring systems support both teachers and students by
combining the benets of individualized delivery and manageabil-
ity by leveraging software personalization, while gamication can
be used to enhance motivation through personalized rewards or
cooperative and competitive activities.
The rest of the paper is structured as follows. Section 2presents
the needed background and the motivations that led us to realize
6
Open Educational Resources are public domain or open licensed educational
resources https://www.unesco.org/en/communication-information/open-solutions/
open-educational-resources
7
https://www.linkedin.com/business/talent/blog/talent-strategy/linkedin-most-in-
demand-hard-and-soft-skills
8
https://www.linkedin.com/business/talent/blog/talent-acquisition/why-shell-
pushes-hard-on-soft-skills
9https://www.eschoolnews.com/2022/07/26/ai-intelligent-bots/
PolyGloT. Section 3introduces its architecture and explains how
it can help teachers create exible learning and eTutoring experi-
ences. AI planning and gamication mechanics needed for adaptive
learning will also be discussed by the end of section 3. Section 4
will describe a scenario where PolyGloT is used to assist teachers
and students in introductory statistics for a data science course.
Lastly, with Section 5we present a glimpse into future directions
and possibilities enabled by the proposed solution.
2 BACKGROUND AND MOTIVATIONS
Adaptive learning technologies have gained traction over the last
decade. Existing solutions have been successful in both domain-
specic [
11
] and institution-wise implementations. In 2015, the
Colorado Technical University (CTU) reported that, following their
Intellipath
™
adoption, the course pass rate rose sharply by 27%,
and the average grade and retention rate were also signicantly
aected. CogBooks
10
conducted a pilot study with Arizona State
University (ASU), resulting in ndings similar to CTU.
Smart Sparrow
11
(recently acquired by the industry-leading Pear-
son) provides a user-friendly WYSIWYG content authoring tool to
create interactive online experiences. Realizeit, Cerego and Cog-
Books create content for their adaptive platforms by partnering
with institutions like ASU, the American Psychological Association,
or directly with industrial partners and customers. Each platform
also provides in-depth analytics on the students’ performance, er-
rors, time spent learning, and other related metrics. Those solutions
solve the problem of individualization and verticality of linear learn-
ing paths but overlook a fundamental factor: familiarity with the
tooling and the environment. Job candidates in software engineer-
ing, for example, are expected to be somewhat comfortable with
industry-leading tools and methodologies. Therefore, not only de-
livering interactive exercises directly on those tools enables richer
experiences but also better prepares the students for their future.
High school electronics students, instead, may benet from hands-
on experience with physical devices like Arduinos
12
. Moreover,
integrating education-ready tools like pi-top kits
13
or Swift Play-
grounds14 allows access to already existing quality resources.
Novel interactive learning experiences can emerge even with
the use of available technologies. Augmented and virtual reality
are still emerging, but education and industry are already taking
advantage of their benets [
8
]. In the classroom, they are used to
promote interactivity and generate engagement. In the industry,
they found use in, among the others, remote maintenance on indus-
trial devices, surgery training programs, aviation, and even military
equipment
15
. Similarly, the use of voice user interfaces (like vir-
tual assistants) may be a means to mimic a study companion that
can ask questions, give feedback and help with misunderstandings.
Adaptive educational technology should allow and encourage the
10https://www.cogbooks.com/
11https://www.smartsparrow.com/
12https://www.arduino.cc/
13
pi-top produces high-quality educational kits for electronics and robotics based on
the Raspberry Pi. https://www.pi-top.com/
14https://www.apple.com/swift/playgrounds/
15
Since 2019 Microsoft and the US Army have collaborated on using HoloLens to
enhance soldiers’ situational awareness https://news.microsoft.com/transform/u-s-
army-to-use-hololens-technology-in-high-tech-headsets-for-soldiers/