Tiago Thompsen Primo

An architecture and algorithm for context-aware resource allocation for digital teaching platforms

Digital Teaching Platforms (DTPs) are aimed to support personalization of classroom education to help optimize the learning process. A trend for research and development exists regarding methods to analyze multimodal data, aiming to infer how students interact with delivered content and understanding student behavior, academic performance, and the way teachers react to student engagement. Existing DTPs can deliver several types of insights, some of which teachers can use to adjust learning activities in real-time. These technologies require a computing infrastructure capable of collecting and analyzing large volumes of data, and, for this, cloud computing is an ideal candidate solution. Nonetheless, preliminary field tests with DTPs demonstrate that applying fully remote services is prohibitive in scenarios with limited bandwidth and a constrained communication infrastructure. Therefore, we propose an architecture for DTPs and an algorithm to promote the adjustable balance between local and federated cloud resources. The solution works by deciding where tasks should be executed, based on resource availability and the quality of insights they may provide to teachers during learning sessions. In this work, we detail the system architecture, describe a proof-of-concept, and discuss the viability of the proposed approach for practical scenarios.

Towards an Educator-Centred Digital Teaching Platform: The Ground Conditions for a Data-Driven Approach

We introduce innovations in a Digital Teaching Platform (DTP) through tools centred on supporting the teacher. We focus on the utilisation of data about the students and the class in order to recommend actions and content for the teacher. For this, we need a platform with novel capabilities. First, we augment the content delivery application with data collecting capabilities. Second, we create a cloud-based analytics engine that infers student profiles and context parameters from multi-modal sources. Third, we provide a web-based platform for content composition that makes use of the inferred student and context profiles to support teachers in lesson planning. Our solution implements the complete cycle from content composition to delivery and adjustment, allowing for the research and development of new features and intelligences in Digital Education.

A Recommender System that Allows Reasoning and Interoperability over Educational Content Metadata

This work presents a recommender system infrastructure for educational material described with metadata. The idea is to provide a set of flexible premises that allows reasoning and improved personalized results in learning content recommendations. For its operation we propose the use of the OBAA standard, which is an extension of IEEE LOM that provides interoperability among hardware platforms contextualized with the Brazilian Educational domain. The technological core of this infrastructure is based on the use of FOAF to describe user profiles, an OWL Ontology to describe specific domain features as well as to facilitate the reasoning process, a Web Service that connects to a federate educational content repository, and a Collaborative Filtering Algorithm as an algorithm recommendation.

A Quantitative Analysis of Learning Objects and Their Metadata in Web Repositories

This work conducts a quantitative analysis of a number of Learning Object Repositories (LORs) of Learning Objects (LOs) in both English and Portuguese languages. The focus of this exercise is to understand how the contributors organize their metadata, the update frequency, and measurement upon LOR items such as: (i) the size distribution; (ii) growth rate, and; (iii) statistics about metadata completion, blank fields and LO types. We conclude our analysis with a discussion about the implications of our findings upon tasks such as LO search and recommendation.

Social Computing in Digital Education

This work presents amodel for the selection of Collaborative Learning (CL) techniques considering specific characteristics and needs of the activity that teachers want to perform within their educational practice. This model considers the representation of the activity in terms of the required competencies defined from Bloom’s taxonomy. Then, using the characterization of a set of techniques conducted by experts, an algorithm is used for providing an affinity measure, doing a recommendation of the technique to use. A validation of the model from three case studies is also described, carried out by comparing experimental and control groups. The results show that CL allows for achieving better academic performance, but also that those techniques proposed by the recommendation model exhibited higher performance.

Exploratory Apprenticeship in the Digital Age with AI Tools

Along this decade, advances from Cognitive and Computing Sciences disturbed the college campus, namely, via the availability of online courses (MOOCs) and tools, and also the secondary schools with e-learning environments. The technological impacts helped to democratize traditional university education and brought, to everywhere in the world, the teaching of wonderful professors through Internet (YouTube). Yet, the lectures are still alive, and the old fashion pedagogy seems in good health, most of the students do not participate fully in the process of learning. What is missing is a profound shake-up of mentalities, new initiatives to accelerate learning research and the discovery of the resilience of the whole process. Artificial intelligence and information and communication technologies (ICTs) in general are sound proposals to open alleys till a happy solution, and in this study, we try to show some ideas about the disruption of the traditional classroom scenario due to the availability of innovative online resources and computers (tablets, smartphones, and laptops).

Towards a Digital Teaching Platform in Brazil: Findings from UX Experiments

This work discusses the usability experiments conducted around a proof-of-concept implementation of a novel tablet-based Digital Teaching Platform (DTP). The platform is intended to address specific issues with tablet usage in a classroom setting, and address problems with technology adoption in education, particularly in Brazil. We evaluated the DTP in two separate studies, a Usability experiment in a laboratory setting, and an in-situ experiment in Brazilian classrooms, with the aim of identifying specific problems with the current solution, and identifying usage patterns that better engage students in the classroom. We found that our DTP leads, overall, to a very satisfactory experience. However, any such platform aimed at classroom usage should take special care to address note-taking, and tasks related to collaboration, sharing, and general social aspects of the classroom experience.

An Analysis of Applying the Short Bridge Method to Digital Education

This work seeks to build a new kind of classroom experience by rethinking how educational content is currently transmitted and consumed at schools. This work presents the results of applying the Short Bridge Method in the education context. We evaluate how this approach contributes to the class composition process by providing tools that support educators and school administrators to plan courses aligned with students’ necessities and learning pace. We present our methodology to identify the work flows and artifacts that impact the class composition activities to support educators and school administrators for personalized learning environments. The results allowed us to understand and map the teacher’s behavior during class preparation and define a set of practices to be incorporate in class composition software.

On the use of inertial sensors and machine learning for automatic recognition of fainting and epileptic seizure

This paper depicts a machine learning method for fainting and epileptic seizures automatic recognition. We evaluated five machine learning techniques in order to find out which classification method maximizes the accuracy level and, at the same time, minimizes the computational complexity since the experimental environment has very limited computational resources (processing power). We prototype such method in a wearable device, taking into account F-Score and Accuracy metrics. The experimental evaluation shows that there are no significant difference between KNN, PART, and C4.5. However, KNN has high computational cost when compared to PART and C4.5. PART has low computational cost when compared to C4.5 since it identified less rules.

Using Semantic Web Technologies to Describe an Educational Domain

This research presents a method to describe Learning Objects as Semantic Web compatible Ontologies. The proposed method divides the Ontologies among three layers. The first is composed by the knowledge domain, the second by the LOs and their relations, and the third is responsible for knowledge inference and reasoning. As study case, it is presented the Ontologies of LOM and OBAA metadata standards as part of the Layer One. The layer two is composed by the description of sample Learning Objects based on the properties and restrictions defines by the layer one ontologies. The layer three describes the knowledge inference axioms, which we defined as Application Profiles. Our current results can be resumed as a contribution to Ontology Engineering for Semantic Web applied to Digital Education.