News: In the Media

João Pedrosa was interviewed by the “90 Segundos de Ciência” Podcast on Antena 1 about CMU Portugal’s Large Scale project, TAMI. The project’s main mission is to develop Artificial Intelligence algorithms to support a medical diagnosis.

TAMI will create a set of tools based on these AI algorithms that will explain to clinicians and researchers the diagnosis of a specific disease and its causes, focusing on cervical cancer, lung diseases, and eye diseases. João Pedrosa, researcher at INESC TEC and Professor at FEUP, explains “the images we use in the project are Chest radiography (CXR) images. Radiologists use these images to screen for various pathologies such as pulmonary nodules, atelectasis, pneumonia, or COVID-19. As several pathologies can be diagnosed from these images, the experience and knowledge of a radiologist are necessary to analyze them.” The idea behind the TAMI project is that automated analysis systems can help radiologists and other clinicians look at these images and make the right decisions for the patient.

Recently the project’s team published a paper that was featured in the media on the use of Artificial Intelligence (AI) and Deep Learning (DL) in the complementary diagnosis of COVID-19 by Chest radiography (CXR) through deep learning. More about this paper here.

The project, being led by First Solutions, with INESC TEC, Fraunhofer Portugal, Administração Regional de Saúde do Norte (ARS Norte), and CMU Electrical and Computer Engineering Department, will work on a commercial, scientific, and academic platform that will provide “consumers” access to results and explanations of diagnostic orders, filtered data sets access for investigators or scientists, and a knowledge base for academic purposes.

More about the TAMI project here.

Available online at 90 Segundos de Ciência website.

CMU Portugal project AGENTS – Automatic generation of humor for social robots – was featured in Público Newspaper. The two-page printed article focused on the study being carried out between May and June at ISCTE (Instituto Universitário de Lisboa), involving around 60 people and two robots, to evaluate how can they perceive and express humor.

The experiment, led by Ph.D. researchers Raquel Oliveira and Inês Batina, invited Portuguese native speakers to play an adapted version of the card game “Cards Against Humanity” with Robots Emys and Glin. Público journalist Teresa Serafim played the game and wrote about her experience with the two robots to share a closer view of how everything worked.

“We sat at a table across from Emys and Glin. Before we started playing, researchers Raquel Oliveira and Inês Batina placed sensors on me. Some electrodes are in the ankles and others under the collarbone. While playing, they measure our heart activity. On the fingers, others will monitor electrodermal or galvanic activity, which assesses the sweat production on the skin and may be related to the emotional activity. Shortly after, the scientists left the room, and I was alone with the robots: the game started! The rules are on the table, but the robots also explain the game. Emys speaks and gives us instructions. On a screen facing us, the names of the participants appear – mine (Teresa), Emys’s, and Glin’s. Then they release black cards with the first words of a sentence that will have to be completed by the three players. I am given some options with text segments on a tablet to fill in those sentences. On the other side, the robots respond as well. At the end of each round – when everyone has chosen an expression, we all voted for the option we found the most fun.”

During the game, the involvement between humans and robots was being evaluated, as well as data that allows analyzing humor in the interaction between humans and robots. This experiment is one of the components of the CMU Portugal project AGENTS, which aims to explore how humor can be used to create more natural and realistic interactions with social robots and virtual agents.

According to Ana Paiva, “the idea is to create a dataset that will allow training a system to identify and generate jokes or humorous sentences that will afterward help robots to generate jokes automatically.” The researcher also explains that the aim is to examine how humor is produced and its intention so that it can be used in long-term therapeutic scenarios.

“The idea is, in fact, to use what we learn here so that we can, in the future, improve the interaction between humans and robots by making them feel good and engaged, making the person want to interact with the robot. If people feel good, it will positively affect their health, namely in the elderly population.” Ana Paiva

The project PI explains that other studies have already been successfully carried out on robots and the elderly, supporting those who need companionship and/or have cognitive problems. The project ends at the end of June. Still, the team has plans to pursue the research: “We have high expectations about the database that we are still collecting and see if we can generate interesting models,” states Patrícia Arriaga.

AGENTS is a collaborative CMU Portugal project between Instituto Superior Técnico, INESC ID, ISCTE, and the Language Technologies Institute at Carnegie Mellon University.

The project is led by Ana Paiva at Técnico and Inesc ID, Patrícia Arriaga at ISCTE and Louis-Philippe Morency at CMU. In addition to Raquel Oliveira and Inês Batina, the team also includes Amirali Bagher Zadeh, João Barreiros, and Rui Prada.

Full article at Público or download pdf.

CMU Portugal National co-director Inês Lynce was part of “Gato de Schrödinger” (Schrödinger’s cat) podcast by Exame Informática magazine.  In a discussion on “Artificial Intelligence – threat or opportunity,” the INESC ID President and professor at Instituto Superior Técnico was one of the two invited speakers, along with João Castro from Nova SBE, a conversation led by journalist Sara Sá.

The main topics of the podcast were the challenges and the potential of Artificial Intelligence (AI). When questioned on how AI can contribute to improving our existence and make our lives better, Inês starts by stating that “we must not take humans out of the loop. There must always be a human looking at things, from health applications to job selection processes. If we hand everything over to machines, it can go wrong. On the other hand, with the current technology and its complexity, it is important to understand what is behind the decisions. This is an important area with a large European investment: an AI that gives explanations.”

Full video available here (in Portuguese).