Bonjour @TransilienSav. J’ai pris le metro de station Belleville à CDG. J’ai acheté un billet pour ~11€. A Chapelle, la machine (pour sortir le metro) a mangé mon billet. J’ai dû payer 50€ à CDG à cause de cette machine défecteuse. Je veux procéder à une réclamation. Merci.
Say HELLO to our #GiftBetter Guide featuring some tastemakers, and business owners just like you.
Make the holidays count for more this season, shop independent with us 💚🛍️ utm.io/ucT84
From @WSJopinion: Don’t be misled by the term "elective." Temporary bans on elective surgery prevent treatment for cancer, heart disease, cataracts and other serious ailments, writes Bob Kerrey. on.wsj.com/2VJvkGh
Solid weekend #PyConCA2019! Met awesome people and listened to excellent talks. Also had the opportunity to present a topic that is a great interest of mine: algorithmic bias in ML. speakerdeck.com/topspinj/alg…@pyconca
Electronic medical records are messy and need to be standardized, cleaned, and wrangled before being fed into a machine learning model. Check out my latest article which shares some techniques and tools on how to handle this sort of data: link.medium.com/qHmFPhybw0
I like reading articles on Medium so thought I'd try writing something myself. Just published a post on the process of building a recommender engine: link.medium.com/SVlj4zsCcZ
I'm doing a short survey on deepfake technology: bit.ly/2XqoKD8. If you have a couple of minutes to provide your input, it would be greatly appreciated 😀
Presented some of my work on building a clinical diagnostic model @pycon yesterday. Slides are posted here: speakerdeck.com/topspinj/how… (warning: there’s a lot of data pre-processing 👍) #PyCon2019
Slides to my #PyDataDC talk "A Brief Introduction to Hyperparameter Optimization" can be found here: bit.ly/2Bf7CZb which focuses on medical data and walks through a case study of building a sepsis prediction model.
Another package to add to my list of optimization tools: TPot! Thank you person in the audience for pointing this out 🙂 github.com/EpistasisLab/tpot…@PyDataDC
Can deep learning–based models be used to improve the diagnosis and clinical management of prostate cancer? In a new paper, we discuss a system that mirrors a pathologist’s workflow, potentially enabling doctors to better match patients to therapy. goo.gl/y1y5wf
Just another example of how a biased dataset can give you unfortunate results. I really hope Malays and Turks don't rely on #googletranslator to learn English...