Fernando De La Torre:

 “As Educators We Teach Dual Degree Students What Top Research is About”

 Fernando De La Torre 2   Fernando De La Torre, a professor and researcher at Carnegie Mellon University (CMU), is deeply involved in the Carnegie Mellon Portugal Program, co-advising the dual degree Ph. D. students Ricardo Cabral and Jayakorn Vongkulbhisal, as well as hosting faculty members from Portuguese universities. In an interview to the CMU Portugal Program, he had nothing but pleasant things to say about the Program.

According to Fernando De La Torre, the CMU Portugal Program “is amazing” and he only wishes he had had the chance to apply to a similar Program when he was doing his Ph.D. The Program “is excellent and it is a win-win situation for everyone: for CMU, for Portugal, for the students and for companies. The best business is when everybody gets something out of it and this is a perfect business idea for several reasons,” the professor confesses.

Fernando De La Torre has been working for the last four years in collaboration with João Paulo Costeira and Alexandre Bernardino, professors at the Instituto Superior Técnico of the Universidade de Lisboa (IST-UL), as part of the CMU Portugal Program. This is a partnership that works at a technical and human level, in the researcher’s opinion, because “we are in the same wavelength in the way we approach many things in research and also in life,” he stresses. “It has been very exciting. I typically go to Lisbon twice a year to see them and the students,” he adds.

A good paper typically takes more than one year of work for the student.

But the greatest advantage of being part of the CMU Portugal Program is, according to Fernando De La Torre, “having access to very talented students like Ricardo Cabral and Jayakorn Vongkulbhisal,” who the professor is currently co-advising with João Paulo Costeira. And the fact that the students receive support from the Program is also a plus “because we are not obligated to have deliverables for projects, and so we are able to focus more on basic science,” he states. “This gives us more flexibility to do basic research, which is a great advantage.”
Because research is important to the students’ development, the advisers’ role as educators is to try to teach them what top research is about. According to Fernando De La Torre, “CMU and IST-UL have high standards for research and a common problem is that students believe they have to publish extensively in order to get a good job and that is not true,” he reveals. In fact, this is a general problem in the scientific community “because some researchers are more worried about de quantity of the papers that they publish than the quality,” he adds. But research is something that requires “a lot of hours, students have to be very meticulous with every sentence they write, and run experiments adequately.”
According to Fernando De La Torre, “a good paper typically takes more than one year of work for the student. It is very uncommon that one student can publish more than one good paper a year as first author. And that is the difference between CMU and IST-UL and all the other universities because our standards for the quality of research are very high. Usually one good paper takes two years,” the professor explains. And he stresses: “these quality papers are the ones that are going to have an impact and give you a good job in academia, industry, or anywhere else. It is better to do few things and do them well. When students are in a top institution they learn that this makes a difference.”
While at CMU, dual degree students are faced with a new reality. “Although for them the language is not an issue, as their level of English is very good, the feeling of living abroad for the first time is not easy, and sometimes they are overwhelmed with the amount of work they have to do,” Fernando De La Torre reveals, especially due to the fact that in the beginning the Program is “very demanding for students because they have a lot of classes, the qualifiers and the Teaching Assistanshipt work.” Another aspect that Fernando De La Torre stresses is the fact that sometimes the Program is difficult for students because they have at least two advisors. For instance, for Ricardo Cabral “it must be very difficult having three advisors because each of us is pulling him in a different direction based on our expertise, but he is managing it very well,” he reveals.
“With this partnership we have produced really high-level and high impact joint publications in top-notch journals and conferences.” This means “the scientific output has been outstanding, comparable to any one of the Ph.D. students here at CMU,” he reveals.
Faculty Exchange Program is a ‘seed’ for collaborations
Fernando De La Torre has also been hosting faculty members from Portuguese universities under the Faculty Exchange Program, in the scope of the CMU Portugal. According to the professor, the collaboration with other faculty members in Portugal through this Program “has also been excellent. In Portugal you have great professionals who are very dedicated to education and with a great scientific background.”
For the faculty members, the Faculty Exchange Program “is very valuable because they see the educational standards we have here at CMU,” Fernando De La Torre states. Even though the CMU professor believes that a four-month period is not much to establish a research collaboration, “for the faculty, the Program has a good educational value because they see how we do things, how we teach at CMU.” While at CMU, the faculty members have regular meetings with Fernando De La Torre “where I talk about how I run my group, the content that we teach and a big difference that we find is that they have to teach many hours in Portugal, and here we have different teaching conditions,” the professor explains, adding that “maybe they can take to Portugal what they learn here and help change the system.”
However, even though it is difficult to establish a research collaboration or produce a research outcome in such a short period, “the contact is very valuable because they get to know the people,” says Fernando De La Torre. And with the new initiatives, more specifically the Entrepreneurial Research Initiatives (ERIs), it will be possible to establish new connections, “which is very valuable. It’s very good to put a ‘seed’ to start collaborations,” he reveals.“ João Paulo Costeira and Alexandre Bernardino and I are also trying to find other funding sources besides CMU Portugal, so I believe this will be a long-lasting partnership.”

About Fernando De La Torre
Fernando De La Torre currently leads two laboratories at CMU. In the Human Sensing Laboratory, he tries to understand human behavior using various sensors, such as audio, video and wearable sensors. “The kind of behavior we try to understand are things like activity, emotion or facial expression. We are very interested in monitoring people, especially for healthcare applications,” the researcher explains. According to Fernando De La Torre, from physiological data, it is possible to understand the activity of people and infer their emotional states. “We have some experience in detecting depression, a work we are developing in collaboration with the University of Pittsburg. The assessment of your level of depression is usually based on the expertise of the clinician and what on how you behave,” the researcher believes. Therefore, what the researchers want to do at the Lab is to provide physiological measurements for depression. By monitoring people, they will try to predict different degrees of depression. With different sensors, the researchers will be able to monitor “hot flashes in women with menopause and men with prostatic cancer, for instance,” he explains.

The other laboratory is the Component Analysis Laboratory, where researchers work on Machine Learning. The goal of the component analysis techniques is to extract useful information from a signal, by decomposing the signal into useful/interesting components. According to Fernando De La Torre, “Component analysis is a very versatile technology that can be used for classification, clustering, visualization or modelling high-dimensional data. It has been applied to solve problems in many scientific disciplines such as signal processing, robotics, bio-informatics, or computer vision,” he concludes.
February 2014