There were a handful of reasons why Google was able to surpass all the other search engines. Google was better than others in *all* of these aspects. …

Given what I know from building search engines and large-scale software systems, here is how I expect the complexity of Google’s search ranking algorithms to have changed over time. The diagram shows long periods of time where the complexity keeps on increasing as you make small incremental improvements to the…

The one way Amazon can win over Google in this space is by partnering better with device manufacturers. To start off, I don’t think the battle between Amazon Echo and Google Home is as important as the one between Amazon Alexa, which is their voice assistance service, and the Google…

At a high level, pointwise, pairwise and listwise approaches differ in how many documents you consider at a time in your loss function when training your model. Pointwise approaches Pointwise approaches look at a single document at a time in the loss function. They essentially take a single document and train a…

Say you want to train a machine learned model to perform a ranking task. A common starter question is: Should I use human relevance judgments or clicks as the target variable for training the model? The answer depends on a bunch of factors such as what you are trying to…

The majority of synonym generation techniques that I have seen in user-facing websites / apps solve it as a two step problem: 1. Candidate Generation At this step, given a word, you generate all possible candidates that might be synonyms for the word. Note that what you mean by “synonyms” usually changes a…

The form factors best suited for the tasks we will do 5–10 years from now — Mobile is eating the world. The current default computing interface for a lot of people, even in markets like India and China, is the mobile phone. Before mobiles, it was the laptop and before laptops it was the PC / desktop. As we fit more and more computing power into…

Machine learning is used at almost every part of the stack at major search engines like Google or Bing. Anything that requires some sort of “intelligence” is often solved using machine learning. Here are some places machine learning is used in search engines: Search ranking Often search engines have multiple phases of…

When is the right time to move to a machine learned search ranker from a simpler score-based one — Traditional Information Retrieval started with score-based ranking for search results using scores such as TF-IDF [1] or BM25 [2]. As search engines got better, the scores they used started getting better. A lot of big search engines use a heuristic / score-based model for search ranking. …