#11 Paris Women in Machine Learning & Data Science: RecSys, Algotrading Contest & Switching Careers to Machine Learning
Despite the cold weather and the snow ❄️, we had 105 attendees for our joint meetup Paris WiMLDS and RecsysFR!
The meetup took place at Criteo Labs, which is part of Criteo, an online advertising company with more than 2,700 employees worldwide. Its core technology lies in machine learning at scale to display personalized ads.
Chloé-Agathe Azencott started the evening by sharing the results of a survey we had launched to better know our community. We are glad to see that we have such a diverse audience, with people coming from many different fields. Feel free to take a look at more details⬇⬇).
Our first talk was by Marie Sacksick about recommendation systems (RecSys) for learning.
Marie is doing her Ph.D. at Domoscio, an edtech company working on adaptive learning. In her talk, she highlighted the differences between the challenges faced when building a classical RecSys versus a RecSys for learning. The three main points addressed were:
- The necessity of user modeling for taking into account user evolution;
- What makes a good recommendation;
- The evaluation of the system is also tricky since it is not obvious to assess if a student has learned a topic.
The third point is interesting from an educational perspective: How can a teacher evaluate the learning impact of a class? Classical methods include exams and tests, but what will the students remember in a month or in a year? Would they be able to apply the lessons learnt in a different setting?
The second talk was about algotrading. Elena Tverdokhlebova, Tatiana Shpakova & Karina Ashurbekova told an inspiring story about winning the International Quant Championship, which was organized by WorldQuant. Before participating in this competition, the trio did not know anything about trading, yet they beat the competition of 10,000 participants 😱! How did they get the idea to do it? After having attended a WiMLDS meetup 💙
The trio started the presentation by introducing basic finance concepts, such as alpha models, then moving to tips & tricks on how to keep improving a model. For instance, the winning solution was a smart aggregation of all the different alphas they developed.
For the final talk, our very own Chiara Biscaro led a panel on “How to plunge into a Machine Learning Career”. Chiara switched from astrophysics to machine learning, and she is frequently asked about the difficulties of doing such a challenging career move.
The panel featured people coming from different backgrounds:
- Anne Sophie Hamy-Petit, Medical Doctor, @Institut Curie (Paris)
- Ivan Lobov, Researcher @CriteoAILab
- Nicolas Ayroles, R&D Deep Learning Engineer @Dental Monitoring
The panelists highlighted the most important skill to learn: coding. Another point emphasized during the panel was the need for data scientists in healthcare, as this field seems to be far behind in machine learning. Indeed, the main struggles are quality, annotation of the data, and problem definition (i.e., what is the question I want to have answered?).
And last but not least, some breaking news! We are proud to announce an Hors-série edition (in French) featuring Nicole El Karoui and Hélène Périvier. We will announce it on Meetup very soon!
If you want to keep posted about our activities, you are welcome to:
📑check our Google spreadsheet if you want to speak 📣, host 💙, or help 🌠
📩send an email to the Paris WiMLDS team to keep in touch >firstname.lastname@example.org
📍join our Slack channel for more discussions about machine learning, data science, and diversity in tech!