Abstract
Kidopi has developed and markets CleverCare (www.clevercare.com.br), a framework for control, management and guidance of patients who needs continuous medical care. In addition to graphical user interfaces and APIs access, CleverCare features a state machine, a dialog management system, Natural Language Processing (NLP) tools, and integration with message services such as Telegram and SMS that allow communication with the patient in a personalized and active way, providing a closer care. CleverCare has been recognized nationally and internationally by awards from institutions such as the United Nations and MIT and has made significant progress in the marketplace. Today, it is present in places that are prominent in innovation such as Hospital Albert Einstein, Unimed BH and Hospital das Clínicas de Ribeirão Preto. In order to improve the technical quality of this solution, Kidopi was able to improve CleverCare's Artificial Intelligence mechanism through the PIPE program (Phase 1) of FAPESP, as well as their use in wearables and prepared the solution for international expansion through to the English and Spanish languages. This technical development in Phase 1 has proven to be strategic and valid. With the results it was possible to establish an agreement and pilot agreement with the Universidad Autónoma de Bucaramanga (UNAB), in Colombia, for the implementation of CleverCare and conduction of joint validation researches. The evolution of Phase 1 also allowed the study of different scenarios of evolution for the decision-making mechanism. With this improvement study it was possible to identify the main strategies for the resolution of specialized content creation bottlenecks. It was observed a potential to be explored in this project, related to the natural language processing that, through the creation of content and corpus annotation, will allow improvements such as the creation of medical protocols based on the most recent publications of renowned journals in the area, intelligence based on literature, increasing the amount of information cured that can solve patients' doubts, and reducing the need for human intervention during dialogues. The goal of this project is to improve CleverCare's natural language processing to extract knowledge from free texts in Portuguese, English and Spanish, allowing a better classification of patient responses to CleverCare questions (for dialog maintenance) as well as answers to user's questions and make the development of new features for applications of interest to the company. With these improvements, it is hoped to make CleverCare more competitive in the national market and increase the patient adherence to his treatment plan and track case outcomes. It's also expected to enable more efficient and fluid dialog management, providing a user-friendly experience. (AU)
| Articles published in Agência FAPESP Newsletter about the research grant: |
| More itemsLess items |
| TITULO |
| Articles published in other media outlets ( ): |
| More itemsLess items |
| VEICULO: TITULO (DATA) |
| VEICULO: TITULO (DATA) |