News: Inside Stories

Luís Borges is a CMU Portugal student in Language Technologies, having started his PhD in 2019. His Master’s degree was focused on the application of Natural Language Processing and Deep Learning techniques on the topic of fake news detection. For his PhD, he is working on Neural Information Retrieval, also leveraging NLP and Deep Learning tools. Currently, Luís is in his first year, doing required coursework and research, under the supervision of Professors Bruno Martins (IST) and Jamie Callan (CMU).

Why did you choose the CMU Portugal Program to pursue your PhD?

While I was doing my MSc, I was in doubt whether after I finished it, I would start working for a company or I would go on to pursue a PhD. Given that my MSc research went really well, it was Professor Bruno Martins, my advisor, who suggested I should consider taking a PhD. After some reflection, I decided to go for the PhD, and started looking for PhD programs that allowed me to spend some time abroad, because besides my technical knowledge, I was also looking to expand my cultural horizons. I also applied for a PhD at IST, for backup reasons, but when I discovered the CMU Portugal program, where I was able to study in one of the world’s greatest universities, and also to spend two years in the United States, I truly stopped caring about any other options.

Were you surprised to be admitted considering the challenging acceptance rate at CMU?

I was in Japan when I knew I had been accepted, and what a celebration it was! I’ll be honest, yes, I was surprised. In such a top university like CMU, with a relatively low acceptance rate, I couldn’t really be surprised if I had been rejected. This essentially serves as motivation, since I know I’ve been given a once in a lifetime opportunity, I must work hard not to waste it.

You’ve just started the Program at IST. How is everything going?

Everything is going great! Coursework is going as expected. Regarding my research, I have weekly meetings with my advisors, via Skype, where we discuss the work I developed during the week, and where I’m given new tasks to implement and work on. Communication between me and my advisors is excellent, and the work I’m developing is challenging, like I was hoping, and expecting.

What are your main expectations for these 5 years?

I essentially expect growth, both on an academic/technical level, and also on a personal level. On a personal level, I’ll be able to spend two years in the U.S, on a different cultural context, and benefit from all the advantages that brings me. Regarding technical knowledge, I will have the opportunity to study and to work with very intelligent people, who are at the top in their research fields, and I hope to learn a lot from that experience. I’m not expecting everything to be a bed of roses – a PhD is a long and arduous journey, which requires high levels of motivation, hard work, and mental resilience, but I know that in the end, this program will be highly beneficial for my career, and I will feel proud and accomplished I was able to finish it.

For those thinking on applying this year what advice would you give?

Go for it! You have nothing to lose. Still, have a backup plan. Do not underestimate the required English tests! If you are accepted, be ready to give everything you have to take the most out of the great opportunity you have been given.

Cátia Baptista is a Human Resources Business Partner at Feedzai, the first CMU Portugal start-up and the market leader in fighting financial crime with AI. The Portuguese tech company is coding the future of commerce with today’s most advanced risk management platform powered by big data and machine learning. Founded and developed by data scientists and aerospace engineers, Feedzai has one mission: to make banking and commerce safe. With more than 500 employees and 60% year-on-year growth, Feedzai is considered to be the best in class by Aite and one of the most successful AI companies by Forbes. The world’s largest banks, processors, and retailers use Feedzai’s fraud prevention and anti-money laundering products to safeguard trillions of dollars and manage risk, while improving customer experience.

Feedzai was founded in 2008  by Paulo Marques and Pedro Bizarro, two faculty members of the Faculdade de Ciências e Tecnologia da Universidade de Coimbra (FCTUC), and Nuno Sebastião, product manager at the European Space Agency (ESA). This innovative startup focused in processing large volumes of data with low-latency, producing actionable information in real-time.

In Portugal, there’s an estimated lack of 15K to 25K engineers and most ICT companies have a difficult time recruiting. As Feedzai Head of HR, what is your view on this situation?

As a global company we see this challenge not being exclusive to our country, rather a tendency of the industry itself. When we take a closer look to Portugal we can see the IT market started to grow fast in the past few years, with new start-ups emerging and other big corporations choosing to open their innovation centers here as part of their strategy, so the hiring demand of this type of talent got even higher which amplified the challenge. Despite of that, Portugal is still in a good place when it comes to the pool of talent it has, namely due to its great higher education system, including in the ICT area.

In your opinion what could be done to change this scenario? 

 Due to its complexity, you would need to approach the scenario from multiple perspectives, ensuring we have more talent prepared in the market with the right competencies for this area. For that to happen, companies and academia should work together and align current and future needs of the industry. While we don’t have enough engineers locally, having more programs/initiatives (like the “Tech Visa”, launched by IAPMEI) that can help streamline the process of hiring highly qualified talent from other countries, would be greatly appreciated.

Feedzai is, since 2018, a CMU Portugal Industry Affiliated partner. Do you consider that establishing partnerships with the academia is one possible way to achieve the change?

As mentioned before, academia and industry should definitely work hand in hand. At Feedzai we work to have different initiatives that can bridge both worlds and strengthen this relationship (and it could not be different when we have two of our co-founders as former University Professors). This relationship is even more important when Feedzai brings added value due to the fact that it is leading one of the most recent areas that emerged in Portugal in recent years: Data Science. A few examples are: having students doing internships with our teams (summer and curricular), having some of our engineers give technical workshops in different Universities, University partnerships that sometimes lead to have our engineers lecturing some of their classes. With this both sides can complement themselves based on their different needs – the industry contributing to promote the development of the competencies that they need; the academia with indicators of market trends and ideas of new potential domains to explore.

Can you share with us Feedzai’s recruitment strategy for ICT professionals?

As the company is experiencing an exponential growth, our strategy is making sure we can attract talent from different sources locally and abroad. Brand awareness is an essential part of the job and for that we have a schedule of the most important events and partnerships to do throughout the year. That includes: attending top University Job Fairs, attending and hosting Tech MeetUps, partner with top Universities and Institutions within our domain.  During such events we don’t only look to talk about what we do, but also how we do it, meaning our culture and principles, as those take also an important role in our recruiting process as well – we want people that like what we do, but that are also aligned with how we do it – Grit, Raise the Bar, Ownership, Think Globally. Looking back, we know that this was one of the key elements that made Feedzai grow the way it has in the past years.

After hiring and since it is a really competitive sector, it’s crucial to keep employees. What is Feedzai strategy to make its employees feel “happy”?

One of the most important things on this matter is to make sure you hear the organization and know that you are tackling the right things in the right moment. That’s why we spend time in hearing what our people have to say about the things we excel as a team and others where improvement needs to happen. We do this through engagement surveys, local (office based) all hands meetings, HR open hours, ask me anything moments with our CEO. All these data points serve as a guideline to define our people’s strategy resting assured that it will have a positive impact in the company and in our people.

The CMU Portugal Program is planning to launch Executive programs in ICT-related areas. Do you think that this could be a way to train and also recruit more people into ICT job positions?

In a time where all the companies are becoming more “tech” (e.g., banks, retail), this is a great initiative to develop managers and above positions on how to lead, manage, support this new pool of talent. Programs like this can help grow and develop important skills and competencies that technical individual contributors are not exposed to on a daily basis due to the technical nature of their work and by doing so it will also have an important impact on their retention.

Rui Correia is a CMU Portugal program Alumnus in Language Technologies who graduated in 2018. He is a Natural Language Processing researcher with interest in applying Machine Learning to real-world problems. He was jointly advised by Maxine Eskenazi at Carnegie Mellon University and Nuno Mamede at IST, with a thesis entitled Automatic Classification of Metadiscourse.

Rui currently works as a Lead Machine Learning Engineer at DefinedCrowd, an intelligent data platform for Artificial Intelligence and Machine Learning that offers efficient data workflows to collect, process and enrich training data by combining crowdsourcing, tools, and machine learning capabilities to accelerate enterprise machine learning training and modeling.

Rui’s self-proclaimed mission (and passion) is to bring together linguistics and technology.

 You are currently a Lead Machine Learning Engineer at DefinedCrowd, can you tell us more about the company and your position?

DefinedCrowd provides a human-in-the-loop platform for data collection and enrichment, which is called Neevo. Through Neevo, we build and enrich a community of contributors who are qualified to work on tasks that may involve text, audio or image, such as Named Entity Tagging, Speech Collections, or Object Detection. At the other end of Neevo, our clients, often other Machine Learning teams, gain access to this pool of contributors allowing them to establish continuous and complex flows of quality annotated data to train their AI solutions.
Our goal in the Machine Learning Team spans across several areas of the platform. From assuring and measuring the quality of the data that is delivered, to optimizing the annotation and generation pipelines, we work in a cycle of research and development to continuously improve the product and guarantee that our contributor’s effort is spent where it is most needed.

In 2018 you finished your PhD. In what ways do you think this PhD impacted your current career?

My PhD allowed me to become an expert in areas I had a real passion for, such as Crowdsourcing and Natural Language Processing. Having this expertise really helped me not to compromise when looking for my next career challenge. I got to apply to companies that were doing things that interested me, and today I get up every day and have the luxury of working in something that I love and that challenges me.

When you decided to pursue a PhD, why did you choose the CMU Portugal Program?

During my Master’s degree at Técnico, I was actually already working on a collaborative project with CMU, which was called REAP – a software for learning a second language through reading real-world articles. The coordinator of such project was Professor Maxine Eskenazi, who would then become my PhD advisor at CMU. This is to say that the decision for me was very natural. Things were going well during the Master’s and I was invited to apply to the program to continue and further develop the work I had already done.

How would you describe your experience, both at CMU and IST? 

Above all, it’s a privilege. The experience of pursuing your PhD in two different institutions, surrounded by two very distinct social and cultural surroundings was incredible. I was able to take relevant complementary courses from the Psychology department while doing my PhD in Language Technologies. How great is that? It was also overwhelming at times and extremely challenging, but in the end, it rewarded me with the greatest sense of accomplishment and pride.

How do you intend to pursue your “self-proclaimed” mission of bringing together linguistics and technology?

For me, one of the most important things in my work is to have a keen sensitivity to the problem I am trying to solve, whether it involves natural languages or not. You may have solved something similar a hundred times before but take your time to understand the current context. What are the specifics and the details that are intrinsic to that problem? Can you take advantage of them to improve your solutions? What new questions can you raise? The bottom line is to know the domain you are working on and stay inquisitive and curious.

For other students that are thinking about doing a CMU Portugal Dual Degree PhD, what would your advice be?

Chose a topic that you love, work hard and master it. Get involved in your research community and take advantage of the fact that you will be exposed to great minds that can give their take on your subject. Other than that, just enjoy the ride. Make friends, go bowling, and get inspired. Go Penguins!

Kazi Mohammed Saidul Huq is the Principal Investigator of the the THz – Communication for Beyond 5G Ultra-fast Networks, one of the 8 Exploratory Research Projects (ERPs) that the CMU Portugal Program has currently running. These projects were selected through a Call for Exploratory Research projects, funded by the Fundação para a Ciência e a Tecnologia (FCT) with the main objective of promoting cutting edge research in ICT.

The project PI is leading the research at the Instituto de Telecomunicações (IT) of Universidade de Aveiro (UA) and in a brief interview explained the goal and focus of this work.

The goal of the THZ project is to “look beyond 5G”. What does this mean and how do you propose to do it?

At the verge of the beyond 5G (B5G) mobile communications era, it is widely agreed that B5G networks should achieve greater system capacity (>100 times compared to 5G networks), higher data rate (in the range of terabits-per-second (Tbps)), and greater user density (to adequately support the Internet of Things (IoT) and Nano-Things paradigms). These enhancements are based on the predicted monthly smartphones traffic that should reach 136 exabytes by 2024, i.e., about 4 times the amount of traffic registered in today’s networks (33 exabytes in 2019). As a result, there is significant interest in the development of innovative solutions for B5G ultra-fast dense heterogeneous networks. It is generally accepted that there are three major ways to obtain several orders of increase in throughput gain: extreme densification of the communication infrastructure, large quantities of newly available spectrum, and massive antenna systems, allowing a throughput gain in the spatial dimension. One of the solutions to fulfill such demanding requirements is to use more bandwidth beyond the microwave (μWave) and millimeter-wave (mmWave) spectra, toward higher frequencies in the Terahertz (THz) frequency range.  THz-band communication is envisioned as a key wireless technology to satisfy real-time traffic demand for beyond 5G systems by diminishing the spectrum scarcity and capacity limitations of current wireless systems.

Briefly, can you explain us what is the main focus of the “THz – Communication for Beyond 5G Ultra-fast Networks” project?

The primary focus of this project is to identify key use cases and frequency bands for THz-based mobile communications; develop new Generalized Hybrid Beamforming for Vehicular Connectivity using THz Massive MIMO and study the impact of 3D channel modeling for ultra-high speed B5G networks that take into account spatial coupling by re-examining legacy engineering tools such as mobility modeling through interdisciplinary design, and involve practical experimentation that will be implemented in system level simulation available in the research group. THz seems to be the key to providing the orders of magnitude of gain required to meet the challenge of this project which will provide the springboard ready to embark on his 5G legacy and beyond.

What are its major challenges?

Although the potential of THz is exciting, the challenges facing a dense urban THz-based cellular network are barely understood, and currently immense. Between rapid urbanization and an accelerating number of cellular-enabled devices per person, the density and physical environment these networks will face is new. Although these trends are becoming more widely recognized, there are few models, measurements, analysis, or evaluation methodologies for a hyper-dense, urban THz cellular system. Considerable basic research is needed to enable the development of these crucial technologies.

Moving toward making THz a reality for broadband communication, this proposal will address a few important issues. These are:

(i) Impact of 3D Channel Modeling for Ultra-High Speed Beyond-5G Networks
(ii) Terahertz Massive MIMO for Beyond-5G Wireless Communication.
(iii) Generalized Hybrid Beamforming for Vehicular Connectivity using THz Massive MIMO

The project was one of the winners of the last CMU Portugal Call for Exploratory projects. In what way was this support important for your research work?

This project worked as a brainchild to arrange a special issue on the prestigious IEEE Communications Magazine (https://ieeexplore.ieee.org/document/8387209) where Kazi Huq played the role of the Lead Guest Editor. Moreover, the finding of this research work will “strengthen” us for attracting new sources of research targeting beyond 5G networks.

You are currently working with Carnegie Mellon University Engineering and Public Policy Department Head, Doug Sicker. How does this collaboration work and what are its main benefits? 

The Carnegie Mellon Portugal exploratory project is an international partnership with the mission to place Portugal at the forefront of innovation in key focused areas of Information and Communication Technologies (ICT). What makes this project so significant is that it gives us the opportunity to work with some of the world’s most brilliant researchers, in this case, Douglas Sicker. By receiving constant feedback from him and in unison with the other researcher, we can gauge the level of our work to the world class level.

By the end of the CMU Portugal grant, what are the expected results?

We are expecting to achieve and publish our results in the high-impact research journals and conference venues. We have already published two journal papers in highly prestigious IEEE and IET journals. It is expected that the results achieved in this project should be able to create synergies with several International institutional and industrial partners and so reinforce the collaborative work which is being pursued, in terms of international collaboration.