An unprecedented opportunity to work on cutting edge research
Author
Prof. Sihem Amer-Yahia
Member of the Steering Committee NRP 75 “Big Data”
Sihem Amer-Yahia was newly elected to the steering committee NRP 75. We asked her a few questions about her research, NRP 75 and Big Data in general.
You are member of NRP 75’s steering committee. What are your expectations towards NRP 75?
NRP 75 recognizes the need to support research in building and improving Big Data infrastructures but also the need to address societal impact by funding two complementary axes: legal questions and application domains. NRP 75 constitutes an unprecedented opportunity for data scientists, domain experts, and law experts to work on cutting edge data-driven research. The expectation for NRP 75 is to continue to fund all three axes and encourage cross-fertilization between axes through sharing data and best practices.
What would be missing if NRP 75 did not exist?
If NRP 75 was not there, researchers and practitioners would continue to get funded solely in their own disciplines. Cross-fertilization that is necessary to the success of Big Data research, will either not happen, or will remain confined to private collaborations and likely not benefit others.
Can you explain to us what Big Data means to you?
The advent of Big Data is putting Data Science at the center of major initiatives in various domains. In academia, an increasing number of Master’s programs are offering a mixed curriculum with a training that combines Data Science disciplines with expertise in a particular domain such as Management and Health. In industry, Big Data is blurring the boundaries between the assignments and responsibilities of project members, and that is a major opportunity in specializing oneself in other domains. As an example, data scientists are developing medical expertise and medical doctors are becoming proficient at using Big Data infrastructures and libraries. Training programs, both in academia and in industry, have to remain aware of this disruption.
Where do you expect the biggest impact? What is nonsense?
The biggest impact of Big Data projects will be attained by projects involving data scientists, domain experts and law professionals. These three facets of Big Data can only be realized with the involvement of researchers and practitioners who are willing to dedicate the necessary time to disciplines outside of their area of expertise. In particular, researchers have the responsibility to advocate in their own communities, research and methodologies from other disciplines. In Computer Science research (data management, data mining, machine learning), we are witnessing the creation of specialized tracks that accept submissions exhibiting mixed expertise in the form of a Use Cases track or an Experimental and Analyses track. That is a good sign and more efforts of that kind should be encouraged to foster cross-discipline collaborations. Nonsense would be to confine the work of each discipline within itself and not listen to the needs and advances in other disciplines.
Let’s talk a little bit about your research: At NRP 75’s Research Day May you talked about “Fairness in Online Labor Markets”. What was your specific research question? Why did you choose this question?
Online job marketplaces are gaining popularity as mediums to hire people to perform certain tasks. These marketplaces include freelancing platforms such as Qapa and MisterTemp’ in France, and TaskRabbit and Fiverr in the USA. On these platforms, individuals can find temporary jobs in the physical world (e.g., looking for a plumber), or in the form of virtual “micro-gigs” such as “help with HTML, JavaScript, CSS, and JQuery”. On these platforms, people are ranked by algorithms according to their qualification for job. The question of studying fairness in online labor markets is essential as more people rely on online platforms to find jobs. My current research addresses the question of fairness in ranking individuals on those platforms and studies which groups are over- or under-exposed to online jobs.
What are the conclusions?
The main findings of my work on fairness in online labor markets is the need to study and compare how people are treated across multiple platforms including dedicated ones as well as Google Search that recently introduced job offers in its organic search results. This kind of work can only be conducted as a collaboration between data scientists and social scientists, in particular, from the legal domain. Several other open questions remain to be solved including which legal framework should be applied to this work and how to repair discrimination.