mattwilson

3 ms·

1. 1.▲
   The Winner's Curse(tech.nextroll.com)

   1 points by mattwilson 6y ago|0 comments
2. 2.▲
   Tips on Remote Work from Remote Engineers(tech.nextroll.com)

   2 points by mattwilson 7y ago|0 comments
3. 3.▲
   Questioning Second-Price Auctions in Ad Tech(tech.adroll.com)

   4 points by mattwilson 8y ago|0 comments
4. 4.▲
   by mattwilson 11y ago

   The Factorization Machine term can be computed in O(kn) time where k is a hyper-parameter of the model and n is the number of features in the model. The features are embedded in a k dimensional space so there is nothing infinite dimensional
5. 5.▲
   by mattwilson 11y ago

   Without factorization machines, you would have O(n^2) quadratic weights w_{i,j} to learn. This is bad in a setting like online advertising where you might have millions of features (i.e. n is large) for two reasons. First, this makes traini