#21 Paris Women in Machine Learning & Data Science Hors-Série n°6: Pi World Record, Computer Vision & Molecular Simulations

📸 Caroline Chavier, Marina Vinyes, Žofia Trsťanová, Despoina Ioannidou, Natalie Cernecka and Emma Haruka Iwao

🔔 On the 8th of October, PwC welcomed us in their beautiful office in Neuilly-sur-Seine. For the sixth time, we organized an exceptional “hors-série” meetup which gathered world-class female speakers.

The meetup’s introduction section was made by our co-founder Caroline Chavier. She welcomed Emma Haruka Iwao with a few words in Japanese.

1️⃣ The France is AI conference happened on the 23rd of October 2019.

2️⃣ We shared an AWS / Reinforcement Learning challenge. You can create a self-driving model using Reinforcement Learning & train on the AWS DeepRacer 3D racing simulator to improve your performance. Register here.

3️ We shared a free workshop about “Introduction to Machine Learning in Python using SciKit Learn”. Register here.

A free workshop that Manon Ansart shared with us

4️⃣ On the 3rd of October, our very own Marina Vinyes This week, our very own got the honor to introduce WiMLDS and WiMLDS Paris to Bruno Le Maire, France’ Minister of Economy and Finance. You can also spot introducing #RecSys to French

🎬 Here is the video of the meetup:

Emma Haruka Iwao

Emma Haruka and her team calculated 31.4 trillion digits of Pi in 2019 and broke the world record in the Pi computation. She discussed the nature of the calculation, the architecture, challenges, and techniques, and of course the history of Pi computation. Calculating Pi has been the Emma’s childhood dream, and she described how her dream grew to the world record.

Slides from Emma Haruka Iwao
Despoina Ioannidou

During her talk, Despoina Ioannidou presented several challenges hat she faced with her team while developing scalable computer vision applications at Meero. She explained how they used Deep Learning models to upgrade a photo shoot picture to a professional look. Aesthetic improvement is challenging because it involves a subjective evaluation. In Meero they rely on their own dataset of pro-edited real estate images. They build their own neural network combining Deep Bilateral Learning for Real-Time Image Enhancement and U-Net.

Slides from Despoina Ioannidou
Žofia Trsťanová

Our last talk was from Žofia Trsťanová, Machine Learning engineer at Criteo. She talked about the work she did during her Ph.D. and PostDoc on molecular simulations and how to accelerate them.

Molecular simulations are useful in numerous fields like chemistry, physics, biology (even awarded with Nobel Price). The idea is to simulate interactions between atoms in order to estimate macroscopic properties which amount to compute high dimensional integrals. The idea is to use Markov Chain Monte Carlo that allows the computation of expected values with respect to a given distribution. The main difficulty relies in generating a Markov Chain. She focused on Langevin dynamics and explained an algorithm with theoretical guarantees that accelerates the computation.

Slides from Žofia Trsťanová

Before ending this post, the WiMLDS Paris team would like to thank our members for the amazing feedback and support!

We are looking forward to keep sharing knowledge and highlight women professionals in Machine Learning and Data Science. Our next meetup will happen on the 28th of November 2019 with Imen El Karoui, Alice Othmani and Sonia Bahri.

If you want to keep posted about our activities, you are welcome to

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

🔗follow our Twitter account, Meetup page, Instagram or LinkedIn page

📍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!

🔥 Feel free to share your company or lab’s job positions for free on WiMLDS’ website.

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