“Estudiamos cómo ayudar a los médicos con la IA”

Publicado en el suplemento

Actions for Citizen Participationen Diario Perfil

Como suele ocurrir con otras questiones, el asunto acá no es la tecnologia en sí misma, sino el uso que se le da. Diego Fernández Slezak is a doctor in Computing Sciences of the UBA, a member of the Instituto de Investigación en Ciencias de la Computing del CONICET and a professor in the Faculty of Exact and Natural Sciences of the University of Buenos Aires. After focusing on the estimation of parameters in complex biological models, he decided to dedicate himself to research on the frontier between Artificial Intelligence (IA) and neuroscience. He is the founder of Entelai, the first company in Argentina with regulation of ANMAT to operate algorithms in medicine. As a result of these investigations, the firm has developed an app in the form of a prototype that, through automatic text processing, collaborates with psychiatrists and psychologists to diagnose cases of schizophrenia based on interviews. They also developed “Pic”, a software for the analysis of medical images that provides an aggregated value for all actors of the health system. En dialog con Actionsel specialist dio cuenta de cómo trabajan con estas tecnologías y el futuro de la IA.

– ¿Cómo se emplea Artificial Intelligence en la salud mental?
Our idea is to try to provide quantitative tools in the world of Artificial Intelligence that allow doctors, psychiatrists in particular, to make decisions based on evidence. La pregunta que nos hacemos es ¿que podemos usar de la AI para ayudar a estos profesionales? El psiquiatra toma sus decisions en funcio de entrevistas que les realizas a los pacientes. Entonces, las analiza de forma qualitativa haciendo un study “intuitive” en funcio de todo lo que se fue informando lo largo de su carrera. Pero no tiene las tools que le permite medir quantitativamente eso que genera con los años de practica médica. One of the objectives that we proposed was to try to find the last aspects that psychiatrists intend to look at in interviews with patients and we could generate quantitative indices that allow them to measure what they were looking for “intuitively”. Concretely, we study how to use Artificial Intelligence to generate indices and numbers that help doctors estimate psychiatric characteristics.

– ¿Para eso desarrollaron una application?

Por ahora es un prototype. No es una app public que se pueda bajar cualquier persona. It is a tool for the doctor in the context of the research project. Cuando se habla de una application, se piensa que va a ser un jueguito de computadora que les va a decir si son bipolares o no, y no es así. It is a research project in which we generate, ultimately, the last metrics for different characteristics of speech that doctors do not know quantitatively and we can do it. Es información flamante que ahora hay que aprender a procesarla. The topic is new both in computing and medicine. The latest new data for diseases that are very difficult to diagnose and monitor with technologies that change day by day.

– Could this technology be used in other areas of medicine?

From Entelai, we are also dedicated to automated image analysis with Artificial Intelligence. We developed “Pic”, a tool that aggregates quantitative and objective information in the world of medical images of magnetic resonance where, until very recently, it was done qualitatively. Gente con muy buen entrenamiento y buenos analistas de imágenes podían extract muy valiosa información, pero hoy la computadora ofrece mucísimo más robusta información sin ningún esfuerzo del ser humano. Este no es un prototype, sino que ya es utilizado por empresas de todo el mundo.

– ¿Hay otros usos que se puedan dar a la Inteligencia Artificial en medicina?

We have another project that applies to medicine in general. A patient who has any symptoms can be guided and derived by Artificial Intelligence, thus generating increased information for the doctor so that he can make his work faster and faster when he receives it. Con esta herramienta se reduce la quantity de tiempo y de consultas.

.

Leave a Comment

Your email address will not be published. Required fields are marked *