"The goal behind the Semantic Web is to transform the existing [World Wide] Web by structuring its data so that they can be easily processed by machines. Our first step is to find keys for identifying the data. For data that have been structured using RDF*, information is expressed in what are known as 'triples': subject–predicate–object. An example is 'person–first name–Danaï' or 'plant–variety–Gariguette.' Then, we can develop ontologies that formally describe the hierarchies and relationships among data. For example, an ontology may specify that a person can be married to another person but not to an animal." With great skill, Danaï Symeonidou teaches us about her research in this exciting scientific field. For her, teaching is a natural way to share knowledge. As early as grade school, she saw herself becoming a teacher one day. As an undergraduate, she studied informatics, which "helps to answer questions across all disciplines." She earned her master's degree via a joint programme, where her coursework was split between Greece and France. It was during her master's internship at the University of Paris-Sud that she was drawn to research. She decided to pursue a PhD and studied the Semantic Web for her thesis. Thanks to her skill set, she was hired by INRA in 2015, after spending just a few months as a post-doc. By obtaining a permanent position, she could pursue a productive career in research, an opportunity that she valued after the stress of being a PhD student.
Big data and the semantic web
The big data found within the Semantic Web are highly heterogeneous and come from diverse sources. As a result, "linking the data together is extremely important because it will allow us to build a 'data web' and to share our knowledge with other research groups," explains Symeonidou. During her PhD, she developed algorithms to identify keys for linking data. Later, at INRA, her early work focused on analysing oenological data, namely information on the levels of different compounds contributing to wine aroma. It was her first time dealing with data from the agricultural sciences. She therefore had to modify her methods to deal with numerical data. She also collaborated with several colleagues, notably statisticians and various researchers who could confirm certain findings based on their knowledge of experimental data. She comments, "I needed to how they did things because we don't necessarily have the same lexicon. It was interesting!" She also discovered that certain types of new findings cannot always be confirmed by experts. They can, however, be verified using similar results obtained from a much larger amount of data. Her research fits squarely within the open science movement. "When we develop algorithms, we generally come up with a tool that we then make available to the scientific community. In my field of research, it is normal and obvious that we should share our work. If we don't make our data or methods accessible, it is impossible to reproduce our work, which can render our research unpublishable."
In my field of research, it is normal and obvious that we should share our work
In 2017, Symeonidou received funding from Agropolis to spend 14 months in Ireland learning about "genetic" algorithms. She explains, "To use an analogy, it's like a geneticist who is trying to create offspring that display a combination of their parents’ best traits. In our approach, we compare and contrast a series of different algorithm combinations to increase algorithm quality, which is scored each time." Equipped with this toolkit, Symeonidou will expand the scope of her research. Starting in September, with the help of a PhD student that she will be co-advising, she will tackle the mountain of data that are recorded daily by the machines of a technical platform that focuses on the relationship between plant phenotype and the environment. She explains, "We want to make new discoveries. For example, which factors can increase production? What can these data tell us about how to best manage, water, and prune plants?" With the help of an intern, she also wants to analyse data from the Pech Rouge Oenological Experimental Station. "We want to encourage people to structure their data, to utilise ontologies,” she comments. In both research situations, the goal is to advance understanding of how climate change impacts plants and plant products.
A passion for teaching
"When pursuing a new project, we try to understand everything that has already been done and to build upon it to arrive at something new, which we then share. For me, teaching, training, and diffusing knowledge are all very important! Serving as a PhD advisor is difficult. You must learn to give good advice. However, I believe that it is a new experience that I will truly enjoy," Symeonidou explains.
At present, she is also helping with the Data Science training programme, which is run by Montpellier SupAgro. This love for teaching likely motivated her to master French. She comments, "At Paris-Sud, I was the first and the last instructor to teach in English. I really pushed to be able to do it. They accepted, but they also asked me to improve my French. I practiced a lot with friends, and, little by little, my French got better." Her language skills were not a problem when she applied to INRA. Symeonidou wrote her application in English and spoke French during the interview. She says, "The selection board saw that I was capable of communicating with my future colleagues and that I spoke both languages comfortably."
The life of a young researcher
Symeonidou recalls that "being a PhD student was hard because of the fierce struggle for fellowships,” but she appreciated having an advisor. After she became a researcher, her life changed dramatically. It is now up to her to advise students and to take advantage of her freedom to develop new tools and professional connections. She thus organised a seminar series; the first seminar has just taken place. She explains, "Seminars are important opportunities for researchers. They allow us to develop collaborations but not just. They expose us to other people's work, encourage scientific discussions, and inspire new research. Ideas lead to projects and joint publications.” At least once a year, Symeonidou presents her work to 300–500 scientists at international scientific conferences. In her field of research, conference proceedings have the same value as publications. She comments, "It is a very inspiring and interesting way to share knowledge. You get up to speed on lots of different topics. I also really enjoy giving presentations." Like all researchers, Symeonidou is constantly searching for funding.
Learning is a constant process. It never ends
Her current philosophy? "If I see something really interesting, I apply for it." When she applied for the position at INRA, she thought that she was too inexperienced, but her advisors encouraged her to try anyway. She comments, "Today they say, 'You see? It worked!' I am very happy!" She then laughs, "My friends are all asking me, 'Do you have kids?' or 'Do you have a house?'" Her life has taken her far away from her family and the place where she grew up in Greece, but she is happy to be at INRA, where she works in a "very energetic” team of young researchers. She also likes the city of Montpellier, which is dynamic but not overwhelming. Plus, the climate reminds her of home. Outside of work, she applies her curiosity to home-improvement projects, woodworking, sewing, painting, and the pursuit of new experiences. She also travels a lot, both for work and for family visits. She remarks, "My brother has a post-doc in Luxembourg, and he is learning French. I told him, 'Learning is a constant process. It never ends.'"
31 years old
- Since 2015: INRA Research Scientist
- 2017–2018: Research trip to the Insight Centre for Data Analytics, University of Cork, Ireland
- 2015: Post-doc, Telecom Paris Tech, France
- 2011–2014: PhD focused on discovering keys for linking Semantic Web data, University of Paris-Sud, France
- 2010–2011: Joint master's programme at the University of Crete (Greece) and the University of Paris-Sud (France)
Teaching experience :
- Co-advisor—Data Science training programme, run by Supagro Montpellier
- Instructor—Semantic Web courses at the University of Paul Valéry and the University of Montpellier
- Teaching assistant—University of Paris-Sud (2011–2014)
Manual activities—woodworking, home improvement, painting, sewing, and others.