#38 Paris Women in Machine Learning & Data Science: consumer feedback analysis, clinical applications, & cost of war

WiMLDS Paris
5 min readMay 4, 2022


After a long hibernation, we met face-to-face again on Tuesday 5 April 2022, in the stylish locations of L’Oreal Beauty Tech Square.

Our co-founder Caroline Chavier introduced the meetup, reminding everybody of our motto “sharing knowledge and care”. She suggested some resources on data and AI, such as a forthcoming book of Cathy O’Neil.

She emphasized that this meetup was special, as it took place during the Russian aggression on Ukraine. At WiMLDS Paris we stand with Ukraine, with Ukrainian women and Ukrainian tech community, and believe it is our duty to help Ukraine in these dark times.

One initiative is Science for Ukraine, which matches Ukrainian researchers and students with open academic positions, scholarships, and grants worldwide.

In France, PAUSE program, which offers hosting for scientists and artists in exile, has launched a special fund for Ukrainian researchers, artists, and culture professionals.

Multiple websites help Ukrainians to find jobs, such as this one that helps specifically Ukrainian women to find jobs in the EU.

This Sifted article lists more ways how the tech community can help Ukraine, including donating, volunteering, and offering jobs and relocation.

Additional resources on how you can help Ukraine can be found here.

For this meetup, we reverted to our traditional format of three talks: one from industry, one from academia, and the third on a broader topic of women, technology, and society.

The first speaker represented our host, L’Oreal: Olesia Khrapunova, Data Scientist at L’Oreal Beauty Tech Accelerator. Her talk was entitled “NLP and Computer Vision Applications in Consumer Feedback Analysis”.

Olesia Khrapunova and the audience

She works in the data science team on a global level, not only France. She presented a project about the consumer loop, which would help employees to understand the customer feedback better and at scale.

About the pure data-science part, they used topic extraction using LDA, topic enrichment, and sentiment analysis, thanks to an in-house model. Olesia explained that they used an interesting graph system to group products together through platforms, which allowed sharing feedback across different but similar products, for instance different color lipsticks. This graph was enriched thanks to the images of the product provided by the platform, to determine if two products were similar. They used a resnet network, pre-trained on ImageNet, and then fine-tuned on their data.

Outside that pure data-science aspect, there were other difficulties. One of them was data collection (of course!). As the products are sold on many platforms and often under different names, there are plenty of way to collect feedback, which might come in multiple languages.

If you want to know more, check out her slides below.

The second speaker was Irene Vignon-Clementel, Senior Researcher at INRIA. Her talk was entitled “Blood flow simulation for clinical applications”.

Irene Vignon-Clementel explains

She presented several projects that she conducted with her team, focusing on blood simulation inside the human body. These simulations can help to predict hearth disease evolution or to define a surgery method to limit the risk of high pressure in the vessels.

As a human body is made of 100.000km (!!) of vessels, the traditional blood simulation with Computational Fluid Dynamic (CFD) cannot be applied directly, as it would be far too time-consuming. Multi-level analyses are therefore made, with simplified models for remote organs and vessels (1D or 0D models), and detailed models close to the zone of interest.

When it comes to machine learning, the medical sector always hits the data quantity problem. The main two issues are:

  • the lack of data, e.g. a cohort of 30 patients is considered as a big dataset.
  • the data measurements,e.g. the blood pressure cannot be measured everywhere at anytime.

Thus, the machine learning done is specific, and must include medical expertise to cope with the small datasets available.

If you want to know more how she tackles the issue for disease progression with patient scans, check out her slides below.

By the way, Irène is looking for a PhD Student! You can reach out to her if you are interested.

The third speaker was extra special. Anastasiia Tryputen is an AI enthusiast, researcher, entrepreneur, and founder of two companies, NeuroHarb and Data unBlocked. She is also a Ukrainian, and we invited her to talk about the Russian war against Ukraine and its impact on the country and its population. Her talk was entitled “Cost of war — employment impact”.

Anastasiia Tryputen and the colours of Ukraine

Anastasiia started by sharing with us some facts from Ukrainian history.

Did you know that there are 846 higher education and research institutions in Ukraine, and that Kyiv-Mohyla Academy is considered the oldest educational institution in Eastern Europe?

Did you know that many well-known tech companies, such as WhatsApp, PayPal, GitLab, Revolut and Grammarly, have Ukrainian co-founders?

The Russian aggression is taking a huge toll on the country and its population. Thousands are killed, wounded, missing, or deported. Critical civilian infrastructure, including academic, medical, and cultural institutions, is destroyed. Over 10 million people, mainly women and children, are either refugees abroad or internally displaced.

Anastasiia mentioned multiple ways we can help Ukraine, such as donating, volunteering, and telling the truth about the war.

She then spoke about her own work building a secure platform to help vulnerable refugees find jobs, information, and support.

If you want to know more, check out her slides below.

Additional resources on how you can help Ukraine can be found here.

Final picture of the speakers and organisers

As an extra bonus, the meetup speakers and participants got T-shirts with our logo, which were sponsored by vpTech in September 2019, before the pandemics. We have more T-shirts in two colours and in different sizes waiting for you: come to our next meetup to get one for yourself 😉.

If you do not want to miss our T-shirts and our events, you can:

🔗 follow us on Twitter, Meetup, and LinkedIn

📑 check our Google spreadsheet if you want to speak 📣, host 💙, help 🌠

📍join our Slack channel for information, discussions, and opportunities

📩 send an email to the Paris WiMLDS team to paris@wimlds.org

🎬 subscribe to our WiMLDS Paris Youtube channel

📸 follow the global WiMLDS on Instagram, LinkedIn, and Facebook

🔥 share your company or lab’s job offers for free on the global WiMLDS’ website.



WiMLDS Paris

WiMLDS Paris is a community of women interested in Machine Learning & Data Science