#33 Paris Women in Machine Learning & Data Science: joint meetup with Bucharest
To continue our 2021 European tour, in April we had another joint meetup, this time with WIMLDS Bucharest.
French and Romanian data science communities are closely connected: many Romanian data scientists have studied or are working in France. Perhaps, this is the reason why this meetup felt like one big family!
To kick off the event Marie Sasksick from Paris chapter introduced the agenda and the chapter, which is the 3rd largest in the world, with over 4000 members.
This was followed by a presentation by Ines Teacă and Claudia Voicilă from Bucharest chapter, which is one of the youngest chapters, started in November 2020.
The first speaker was Andrada Olteanu, Data Scientist at Endava. Her presentation was entitled “RAPIDS: erase the waiting hassle” .
RAPIDS is a suite of open-source software libraries to execute end-to-end data science pipelines on GPU developed by Nvidia. The whole point of using GPUs rather than CPUs is that they’re much faster for matrix operations, and ML/DS happens to have a lot of matrix operations. It can be used instead of pandas, numpy or scikit-learn, but you don’t have to learn a new library: the functions are called the same! Isn’t it great?
She also shared a helpful tutorial for those who want to install RAPIDS on Ubuntu.
If you want to know more, check out her slides below.
The second speaker was Ana-Maria Niculescu, Data Scientist at Numeract . Her presentation was entitled “Solutions for running R on AWS Lambda” .
Yes, you might be willing to run R in production! Although python is the most used language in data science, some libraries exist only in R. If you need them, you can choose to translate them with all the maintenance it entails (i.e., not efficient), or you can build your app in R. Ana-Maria went through this process, explaining how we can use R in production thanks to an AWS Lambda.
If you want to know more, check out her slides below.
The third speaker was Denisa Banulescu-Radu, Maître de Conférences (Associate Professor) at the Université d’Orléans (Laboratoire d’Economie d’Orléans). Her presentation was entitled “Data Science for Financial Fraud Detection”.
Fraud can be defined as “an uncommon, cell-considered, imperceptibly concealed, time-evolving, and often carefully organized crime that appears in many types of form”. This definition spells out why fraud is not easy to detect.
During the detection process, you might encounter four main challenges. First, as it’s time-evolving and intentional, you will have to adapt your models often. Second, fraud is rare so here comes the challenge of imbalanced datasets. Third, the performance of fraud detection models is difficult to evaluate. Fourth, most fraud detection models are difficult to interpret.
During her talk, Denisa gave us some solutions for each of these challenges. If you want to know more, check out her slides below.
🎬 You can watch the entire meetup here:
All our meetups in 2021 are joint European meetups! If you do not want to miss them, 🔗 follow our Twitter account, Meetup page, and LinkedIn page.
You can also:
📑 check our Google spreadsheet if you want to speak 📣, host 💙, help 🌠
📍join our Slack channel for more discussions about machine learning, data science, and diversity in tech!
📩 send an email to the Paris WiMLDS team to keep in touch >paris@wimlds.org
🎬 follow our WiMLDS Paris Youtube channel
📸 WiMLDS has an Instagram account, a global LinkedIn page and a Facebook page!
🔥 Share your company or lab’s job positions for free on WiMLDS’ website.
