#16 Paris Women in Machine Learning & Data Science: Graphical Models, Career in Tech & Neural Networks on Device with Greengrass
Our 16th meetup on the 4rth of June 2019 was a special one. First of all, we were hosted by the friendly Doctolib Nerds in their offices, next to the dreamy Monceau park in central Paris. Second, we joined forces with 3 other Parisian women meetup groups 🤩
🛫 We kicked-off the night with a presentation by Marina Vinyes from Criteo Labs, and a WiMLDS Paris co-organizer, about “Learning the structure of Gaussian Graphical models with unobserved variables”. Marina started by introducing graphical models: a graphical model is a statistical model that is associated with a graph where:
- nodes correspond to variables of interest 👩🏻💻
- edges of the graph reflect statistical dependencies between variables. More precisely, edges represent allowed conditional dependencies among the variables 👩🏻💻
She focused on the problem of structure learning where the goal is to estimate the graph (i.e. find the edges) underlying a model. For Gaussian variables and undirected graphical models this problem is slightly easier (because there is a correspondence between the precision matrix and the edge structure) and can be solved with the graphical lasso.
In the last part of her talk, she presented a formulation to be able to learn the structure of the graph in the presence of unobserved variables. This problem is challenging as not only we need to find the edges between observed variables but also:
- the number of unobserved variables,
- the edges between unobserved and observed variables.
Because we are on the mission to encourage women to present their work publicly, it is important that we show the way. The WiMLDS Paris team wanted Marina to remember her first public talk, and to thank her for all the work she does for the meetup! 💐
We were extremely lucky to listen to an inspirational talk from Melanie Warrick, Senior Developer Relations Engineer at Google Cloud. Melanie is co-hosting the Google Cloud Platform podcast: listen, for example, how she explains the difference between ML and AI here 🎧
Melanie shared 7 pieces of career advice and Marie Langé tweeted the main takeaways:
🤹🏻♀️ 1 : BE FOOLISH! Take risks! You have to fail to learn things. Whatever your age, force yourself to do new things to keep learning.
🙏🏿 2. SEEK & GIVE HELP! You have something to share with others, and many people are willing to help you.
🙌🏾 3. PRIORITIZE YOU! You want to make sure you do things you love. Don’t sacrifice your company but make sure you work on what matters to you.
📣 4. ASK FOR MORE! More responsibility, more money, new missions, new opportunities, new things to learn.
✔️ 5. MAKE YOUR PATH! Your do own path, don’t want to be this other person, be you, do what matters to you the way you want to do it.
🔝 6. RESPECT! You need to respect people and be respected.
😎 7. HAVE FUN! Whatever fun means to you. I you love dragons, then make presentations with dragons! But don’t forget to SHOW YOUR WORK on Github.
Melanie shared useful content in her slides. Check it out ⬇️
We ended the meetup with a lively presentation from Sameh Ben Fredj, Data Scientist & IoT consultant Xebia. She shared precious secrets from her work on a playful topic: “a connected bar”! Through her talk, she explained the concept of Edge Computing and how to easily deploy machine learning on devices using cloud services such as Greengrass.
If you want more information, there is also a French presentation 🍹 🍾
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 🌠
📍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 >email@example.com
🔥 Feel free to share your company or lab’s job positions for free on WiMLDS’ website.