Welcome to WebmasterWorld Guest from 54.91.71.108

Forum Moderators: open

Message Too Old, No Replies

Pinterest "Pinnability" Machine Learning For Home Feed Relevance

     
5:20 pm on Mar 22, 2015 (gmt 0)

Administrator from GB 

WebmasterWorld Administrator engine is a WebmasterWorld Top Contributor of All Time 10+ Year Member Top Contributors Of The Month Best Post Of The Month

joined:May 9, 2000
posts:26109
votes: 943


Pinterest engineering has gone into detailed explanation about how it can bring personalized and relevant Pins to the newsfeed.

Pinnability is machine learning it uses to achieve this, and it's explanation is worth reading, especially for those that use Pinterest in our marketing.

In order to accurately predict how likely a Pinner will interact with a Pin, we applied state-of-the-art machine learning models including Logistic Regression (LR), Support Vector Machines (SVM), Gradient Boosted Decision Trees (GBDT) and Convolutional Neural Networks (CNN). We extracted and tested thousands of textual and visual features that are useful for accurate prediction of the relevance score. Before we launch a model for an online A/B experiment, we thoroughly evaluate its offline performance based on historical data. Pinterest "Pinability" Machine Learning For Home Feed Relevance [engineering.pinterest.com]


In training Pinnability models, we use Area Under the ROC Curve (AUC) as our main offline evaluation metric, along with r-square and root mean squared error. We optimized for AUC not only because it is a widely used metric in similar prediction systems, but also because we’ve observed strong positive correlation between the AUC gain from offline testing and an increase in Pinner engagement in online A/B experiment. Our p
12:47 pm on Mar 23, 2015 (gmt 0)

Senior Member

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month

joined:Dec 27, 2004
posts:1987
votes: 73


In order to accurately predict how likely a Pinner will interact with a Pin.

AHA, they are taking about the product image that took. I spent 1200 on camera 100 on the product 2500 on Photoshop Suite.

Why do we promote these scrapers by talking about them, again, and again? They outrank our own work in SERP, yet we give them a word of mentioned?
3:20 pm on Mar 23, 2015 (gmt 0)

Senior Member

WebmasterWorld Senior Member 10+ Year Member

joined:May 14, 2006
posts:692
votes: 59


Why do we promote these scrapers by talking about them, again, and again? They outrank our own work in SERP, yet we give them a word of mentioned?


And they rank pretty well on Google. Google is promoting stolen content these days. Do no evil ...
9:06 am on Mar 24, 2015 (gmt 0)

Senior Member

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month

joined:Oct 4, 2001
posts: 1277
votes: 17


Thanks for the link. This kind of thing should be interesting to everyone here, these are the kinds of algorithms that are increasingly going to organize all of the information we consume. Which is to say algorithms that do independent calculation but are heavily weighted by crowd interaction.
2:01 am on Mar 25, 2015 (gmt 0)

Senior Member

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month

joined:Dec 27, 2004
posts:1987
votes: 73


Innovation that is based on algorithms that is eventually driven by corporate profits. We have seen that before, or maybe not.
 

Join The Conversation

Moderators and Top Contributors

Hot Threads This Week

Featured Threads

Free SEO Tools

Hire Expert Members