150 successful machine learning models: 6 lessons learned at Booking.com – the morning paper


30 bookmarks. First posted by pskomoroch 16 days ago.


150 successful machine learning models: 6 lessons learned at Booking.com Bernadi et al., KDD'19 Here’s a paper that will reward careful study for many organisations. We’ve previously looked at the deep penetration of machine learning models in the product stacks of leading companies, and also some of the pre-requisites for being successful with it. Today’s…
machine-learning  ai  booking.com 
2 days ago by arobinski
Bernadi et al., KDD'19 Here’s a paper that will reward careful study for many organisations. We’ve previously looked at the deep penetration of machine learning models in the product stacks of leading companies, and also some of the pre-requisites for being successful with it. Today’s…
machine-learning 
3 days ago by zdrazil
acolyer's synopsis of the original paper at

https://www.kdd.org/kdd2019/accepted-papers/view/150-successful-machine-learning-models-6-lessons-learned-at-booking.com


1. Projects introducing machine learned models deliver strong business value.
2. Model performance is not the same as business performance.
3. Be clear about the problem you’re trying to solve.
4. Prediction serving latency matters.
5. Get early feedback on model quality.
6. Test the business impact of your models using randomized controlled trials (follows from #2).
ml  models  lessons 
6 days ago by drmeme
Here’s a paper that will reward careful study for many organisations. We’ve previously looked at the deep penetration of machine learning models in the product stacks of leading companies, and also some of the pre-requisites for being successful with it. via Pocket
IFTTT  Pocket 
13 days ago by roolio
150 successful machine learning models: 6 lessons learned at
from twitter
14 days ago by peba
Good tips for real-world production ML/classification adoption.
One tactic Booking.com have successfully deployed in these situations with respect to binary classifiers is to look at the distribution of responses generated by the model. “Smooth bimodal distributions with one clear stable point are signs of a model that successfully distinguishes two classes.” Other shapes (see figure below) can be indicative of a model that is struggling.


Also very interesting to note that people found an over-accurate prediction engine to be "creepy" and an example of the "uncanny valley" effect.
learning  ml  ai  machine-learning  production  booking.com 
14 days ago by jm
150 successful machine learning models: 6 lessons learned at Booking.com Bernadi et al., KDD’19 Here’s a paper that will reward careful study for many…
from instapaper
15 days ago by rogerhsueh
150 successful machine learning models: 6 lessons learned at Booking.com via Instapaper https://blog.acolyer.org/2019/10/07/150-successful-machine-learning-models/
15 days ago by ravivyas
150 successful machine learning models: 6 lessons learned at Booking.com Bernadi et al., KDD'19 Here’s a paper that will reward careful study for many organisations. We’ve previously looked at the deep penetration of machine learning models in the product stacks of leading companies, and also some of the pre-requisites for being successful with it. Today’s…
16 days ago by parsoj
150 successful machine learning models: 6 lessons learned at Booking.com Here’s a paper that will reward careful study for many organisations. We’ve previously looked at the deep penetration of machine learning models in the product stacks of leading companies, and also some of the pre-requisites for being successful with it.
IFTTT  Pocket 
16 days ago by LaptopHeaven