AIX – UX Heuristics for Deep Learning powered Mobile Applications with Image Classification

19/07/2023 06:14

Advances in mobile applications providing image classification enabled by Deep Learning require innovative User Experience solutions in order to assure their adequate use by users. To aid the design process, usability heuristics are typically customized for a specific kind of application. Therefore, based on a literature review and analyzing existing mobile applications with image classification, we propose an initial set of AIX heuristics for Deep Learning powered mobile applications with image classification decomposed into a checklist. These results of this research can be used to guide the design of the interfaces of such applications as well as support the conduction of heuristic evaluations supporting the development of image classification apps that people can understand, trust, and can engage with effectively.

Online Course
In order to facilitate the usage of the checklist we also developed an online course presenting the concepts and heuristics available online for free in Brazilian Portuguese.

Online tool
In to facilitate evaluations using the proposed heuristics and checklist, we also developed a web tool that guides the evaluation by presenting each item of the checklist by its name, brief explanation, image with example and its corresponding response alternatives. In the end, the tool summarizes the results by presenting a list of the checklist items and visually indicating the items that are not satisfied, as well as general percentage of the items that are satisfied. The report can also be downloaded in pdf format. The tool has been developed in javascript and is available online for free in English and Brazilian Portuguese.

More information

C. Gresse von Wangenheim, G. Dirschschnabel. UX Heuristics and Checklist for Deep Learning powered Mobile Applications with Image ClassificationarXiv:2307.05513 [cs.HC], 2023.

G. Dirschnabel. Desenvolvimento de um checklist de avaliação de heurísticas de experiência de usuário de aplicativos com Inteligência Artificial. 2023. Trabalho de Conclusão de Curso. (Graduação em Ciência da Computação) – Universidade Federal de Santa Catarina.

Tags: AIXDeep LearningImage ClassificationMobile applicationUser experience

App in the App of the Season Appathon

07/06/2023 10:10

App desenvolvido pela equipe da iniciativa  como caso exemplo de desenvolvimento de apps com Inteligência Artificial para classificação de imagens ganha menção honrosa no App of the Season Appathon organizado pela App Inventor Foundation/EUA.

App developed by the team of the initiative Computação na Escola/INCoD/INE/UFSC as an example of developing apps with Artificial Intelligence for image classification wins honorable mention in the App of the Season Appathon organized by the App Inventor Foundation/USA.
The app was developed in cooperation with LTH/UFSC and CRLAB/LACEN/SC.

Mais informações

Tags: app inventorAppathonDeep LearningDengueGoogle Teachable Machine

Best research article at EduComp 2023

02/05/2023 15:06

Congratulations to Marcelo and all the co-authors for the article awarded as the best research article by EduComp 2023:

Marcelo F. Rauber, Christiane Gresse von Wangenheim, Adriano F. Borgatto, Ramon M. Martins. Análise do desempenho de aprendizagem de Machine Learning na Educação Básica aplicando a Teoria de Resposta ao Item. In: Anais do III Simpósio Brasileiro de Educação em Computação, online, Brasil, 2023.

The article presents results of research carried out as part of Marcelo’s PhD at PPGCC/INE/UFSC in the context of the initiative Computação na Escola/INCoD/INE/UFSC.

Tags: AssessmentItem Response TheoryK-12Machine Learning

Teaching the development of mobile applications using Artificial Intelligence to underprivileged youth in Brazil

08/03/2023 12:13

As part of the innovative technology education program Pode Crer at the Vilson Groh Institute more than 175 students were introduced to the world of Artificial Intelligence, and learned basic ML concepts in a practical way. In addition to creating their own ML models, they were able to develop intelligent applications to solve problems creatively and critically discuss ethical issues and their impacts, strengthening their active participation in a world increasingly influenced by Artificial Intelligence. 

Course Duration Content Sustainable development goals Stage of learning cycle Tools
Courso 1: Machine Learning (ML) for All! 8h Basic concepts of ML, Development of a ML model for image classification (deep learning), Ethics and social impact Responsible consumption and production: Recyclable waste Develop a pre-defined ML model Google Teachable Machine
Course 2:  Selfie App 4h Development of an app with App Inventor, User interface design, Algorithm and programming Develop a pre-defined app App Inventor
Course 3: Machine Learning in mobile apps 4h Deployment of an ML model in a mobile application in App Inventor Responsible consumption and production: Recyclable waste Deploy a ML model in an app App Inventor/TMIC, Google Teachable Machine
Course 4: Create your own intelligent app (advanced level) 24h App development process, Design thinking, Machine learning, User interface design, Algorithm and programming Related to any of the goals Create a new intelligent app App Inventor/TMIC, Google Teachable Machine

Courses 1 – 3 are available for free in Brazilian Portuguese.

As a result students learned how to develop their own intelligent applications and critically discuss the impacts of Artificial Intelligence within their reality. Observing the interest and dedication of the participants and seeing the results achieved in such a short time was impressive, and demonstrates the importance of such an innovative technology education program, preparing the students to actively participate in a world increasingly influenced by Artificial Intelligence, and inspiring them to pursue a career in this field with so many opportunities. 

Examples of the App Inventor apps created by the students using image recognition for the classification of organic waste, colors for colorblind users and toxic plants.

The App Inventor code of the created apps is available online.

The courses have been offered by the initiative Computing in Schools of the Department of Informatics and Statistics at the Federal University of Santa Catarina that is dedicated to increasing computing education in Brazilian middle and high schools.

As part of the courses students also visited the Informatics and Statistics Department at the Federal University of Santa Catarina. During the visit, the students learned about the undergraduate computing courses and research being carried out using Machine Learning, also as part of the Ampera Racing project detecting objects by an autonomous racing car.

The courses are part of the Pode Crer program, a new education program that blends technology, citizenship, and communication for youth aged 11 – 24 through the development of leadership, creativity, critical thinking and problem-solving in the challenging social context of Brazilian marginalized communities. Pode Crer is free and open for all youth from low-income families, currently taking place in the greater Florianópolis region in the south of Brazil at the Vilson Groh Institute.

The Vilson Groh Institute is a Brazilian nonprofit that focuses on social justice and uplifting marginalized communities in Brazil through education. The purpose of the Vilson Groh network is to open up paths for marginalized youth to take the lead in their own lives and have the opportunity to access universities and the job market, to fight for their rights, to live a life with dignity and to enable youth to dream and have these dreams become reality.

Tags: amperaracingappinventorcneufsccomputacaonaescolagoogleteachablemachineincodineufsc

2022 Teacher Excellence Award Winner da App Inventor Foundation

12/09/2022 15:21

Christiane Gresse von Wangenheim foi premiada com o 2022 Teacher Excellence Award da App Inventor Foundation/EUA por suas contribuições para promover empoderamento tecnológico, criatividade e ação computacional através do ensino do desenvolvimento de apps com App Inventor, do Massachusetts Institute of Technology (MIT)/EUA.

Christiane Gresse von Wangenheim coordena a iniciativa Computação na Escola do INCoD – Instituto Nacional para Convergência Digital do Departamento de Informática e Estatística da Universidade Federal de Santa Catarina.

A iniciativa visa a popularização do conhecimento de computação iniciando na Educação Básica por meio do desenvolvimento de aplicativos que fazem uma diferença positiva, significativa, que tornam o mundo um lugar melhor por meio da busca de um bem coletivo e comum. Com esse objetivo a iniciativa realiza pesquisa, desenvolvimento e aplicação de ensino de conceitos de algoritmos e programação, engenharia de software, design thinking, design de interface e inteligência artificial/machine learning para Educação Básica e Superior.

Conheça os nossos cursos gratuitos de criação de aplicativos móveis com App Inventor!

Esse prêmio também reconhece os esforços de todos os colegas e estudantes que contribuíram para a nossa pesquisa, desenvolvimento e aplicação do ensino de computação e também a todos os participantes dos nossos cursos!

Tags: app inventorcomputação na escola