Profile

Cover photo
49 followers|13,708 views
AboutPostsPhotos

Stream

In the Googleplex

Shared publicly  - 
 
Google Research developed an aLOLgorithm, “Quantifying comedy on YouTube: why the number of o’s in your LOL matter” to measure YouTube videos’ hilarity. Let’s just refer to it as the LOLgorithm, for my ease of typing. Initially, I thought it was a prior year’s April Fool’s Day post. It isn’t! Google began by identifying the humorous videos, which is easier said than done.  YouTube’s search engine is not the greatest. 

Google started with the semantic meaning of the title, designated by the uploader, and the video description and tags if provided. Next, they used viewer reactions as indicated by comments to categorize the humor videos into sub-genre. 

Viewers emphasize their reaction to funny videos in several ways: capitalization (LOL), elongation (loooooool), repetition (lolololol), exclamation (lolllll!!!!!), and combinations thereof. A “loooooool” indicates greater viewer amusement than a “loool”. The final step was ranking the selected videos by relative funniness. The LOLgorithm seems accurate to me. 

An amusing example of contextual/semantic mismatch is an appalling spelling error in a cover of AC DC’s Thunderstruck, performed by The Vitamin String Quartet. The title is listed as TUNDERSTRUK. Looks like the LOLgorithm is working, because that’s what I’m doing now...
YouTube is something of a cesspool, with pockets of exceptional quality here and there. Even the higher quality videos have an ephemeral aspect, mysteriously vanishing or being marked Private, from...
1
Add a comment...

In the Googleplex

Shared publicly  - 
 
I miss Google Sets and Google Square, discontinued as of 2010 or 2011.  This post reminded me of both.
 
Smart Autofill - Harnessing the predictive power of Machine Learning in Google Sheets

Much of Google’s work on language, speech, translation, and visual processing relies on machine learning, where we construct and apply learning algorithms that make use of labeled data in order to make predictions for new data. What if you could leverage machine learning algorithms to learn patterns in your spreadsheet data, automatically build a model, and infer unknown values?

You can now use machine learning to make predictions in Google Sheets with the newly launched Smart Autofill Add-on (http://goo.gl/dghCQs). Smart Autofill uses Google's cloud-based machine learning service Prediction API (http://goo.gl/WmHyAa), which trains several linear as well as non-linear classification and regression models to predict the missing values of a partially filled column in your spreadsheet by using the data of other related columns.

For more information and a tutorial, head over to the Google Research Blog.
1 comment on original post
1
Add a comment...

In the Googleplex

Shared publicly  - 
 
Timely!
 
Google Flu Trends gets a brand new engine

We’re launching a new Google Flu Trends (http://goo.gl/pWqlJo) model in the U.S. that—like many of the best performing methods in the literature—takes official CDC flu data into account as the 2014/2015 flu season progresses. We hope it can help alert health professionals to outbreaks early, and in areas without traditional monitoring, and give us all better odds against the flu. Learn more on the Google Research blog, linked below.
3 comments on original post
1
Add a comment...

In the Googleplex

Shared publicly  - 
 
Very fine post via The Source!
 
All the News that’s Fit to Read: : A Study of Social Annotations for News Reading

News is one of the most important parts of our collective information diet, and like any other activity on the Web, online news reading is fast becoming a social experience. With news article recommendations and endorsements coming from a combination of computers and algorithms, companies that publish and aggregate content, friends and even complete strangers, how do these social annotations affect users' selections of what to read?

Head over to the Google Research Blog, where Stanford University Ph.D candidate and former Google Intern +Chinmay Kulkarni and Google Research Scientist +Ed Chi report on results that suggest that social annotations, which have so far been considered as a generic simple method to increase user engagement, are not simple at all.
1 comment on original post
1
Add a comment...

In the Googleplex

Shared publicly  - 
 
Somehow... #android  was there too. Lower left corner, the little green guy!
1
Add a comment...

In the Googleplex

Shared publicly  - 
 
Last one, merely a month late!
This is an especially short post, as it is a high-level summary of an even higher level summary. Of course, we all know how meaningful THAT is Zeitgeist is a “borrowed word”, from an En...
1
Add a comment...
Have them in circles
49 people
Funny Songs and Masti Maza's profile photo
KatieKoenig BeadsHouston's profile photo
seir marcos's profile photo
Jaimy Zwart's profile photo
Kelly Gibson's profile photo
Alif Aziz Gup's profile photo
Currie Russell's profile photo
terry O'Leary's profile photo
Daniel Bonato Quintanilha (Bona)'s profile photo

In the Googleplex

Shared publicly  - 
 
Ultimate context-appropriate CAPTCHA: The answer is Ohm denominated!
 
The  Electronic Library website of the Moscow Institute of Physics and Technology (http://goo.gl/bb3Ku0) has a CAPTCHA that is a little more challenging than most. Care to take a crack at it?  Or you could just cross that bridge when you get to it...
17 comments on original post
1
Add a comment...
 
Chickless in Seattle with Microsoft  #bing  search
1
Add a comment...

In the Googleplex

Shared publicly  - 
 
Google recently updated its References for Webmasters.
1
Add a comment...

In the Googleplex

Shared publicly  - 
 
Grace Hopper invented #cobol  (among other things). COBOL lives on, keeping her memory and accomplishments alive for us. This is one of my favorite Google Doodles. It is perfect; every detail is correct, and beautiful. Thank you, +Research at Google.
 
How many women can you name who have both a supercomputer and a U.S Navy destroyer named after them? Grace Hopper—who we’re celebrating with a doodle today in the U.S.—is one. “Amazing Grace”’s contributions to computer science made her a pioneer in the field. She created the first compiler for a programming language and led the development of COBOL, the first modern programming language. Happy 107th birthday to Grace Hopper!
40 comments on original post
1
Add a comment...
 
Google Research goes beyond n-grams! Last year, the Google n-gram repository was finally updated with version 2 data, in July. Now they are more publicly making use of their 2010 acquisition of +Freebase and the semantic web-y Internet of Things. I am curious if Google has also integrated #needlebase , a very fine online tool for natural language processing-related (and other sorts of) data analysis. Needlebase was a Google acquisition in 2010 or so, and went offline in July 2012. I have an account on Freebase, but have not found it easy to use. I am very, very curious how Google is able to extract meaningful information from it, as it was kind of a mess the last time I checked. It is indeed vast, and comprehensive, but it requires lots of human organizing, or so it seemed to me. I love ngrams, or for my purposes which are merely casual, recreational NLP
 
11 Billion Clues in 800 Million Documents: A Web Research Corpus Annotated with Freebase Concepts
Posted by +Dave Orr, +Amar Subramanya, +Evgeniy Gabrilovich, and +Michael Ringgaard, Google Research

“I assume that by knowing the truth you mean knowing things as they really are.” - Plato

When you type in a search query -- perhaps Plato -- are you interested in the string of letters you typed? Or the concept or entity represented by that string? But knowing that the string represents something real and meaningful only gets you so far in computational linguistics or information retrieval -- you have to know what the string actually refers to.

The Knowledge Graph (http://goo.gl/sLsqp) and Freebase (http://goo.gl/h3sr) are databases of things, not strings, and references to them let you operate in the realm of concepts and entities rather than strings and n-grams. We’ve previously released data to help with disambiguation (http://goo.gl/20wNW) and recently awarded $1.2M in research grants to work on related problems (http://goo.gl/rheXN).

Today we’re taking another step: releasing data consisting of nearly 800 million documents automatically annotated with over 11 billion references to Freebase entities. To learn more details, and to download the data, visit the Google Research Blog, linked below.
View original post
1
Add a comment...

In the Googleplex

Shared publicly  - 
 
Very BIG data plus Wikipedia via Research at Google: Does more data compensate for idiomatic and contextual complexity of natural language? Maybe.
 
Collaborating with UMass Amherst researchers Sameer Singh and Andrew McCallum, Google Research has released the Wikilinks Corpus, consisting of 40 million disambiguated mentions within roughly 10 million web pages.  By compiling the associations of unique wikipedia URLs (entities) linked to by the hypertext (anchors) of weblinks, the Wikilinks Corpus can help computers with the task of disambiguation (if someone says Stanford, are they referring to a university, a city, or a person?) -- something humans do incredibly well.

To read more about how to obtain the data, and ideas for what you can do with it, head over to the Research Blog post linked below
1 comment on original post
1
Add a comment...
Story
Tagline
Not-affiliated with Google fun and hobby blog about Google
Introduction
In the not-affiliated-with-Google Googleplex is my hobby blog about Google products, news, humor, trivia, images, history. And infrequently, some serious items like internet policy and GOOG Inc updates for investors.