A Tale of Two Bells

An image search on Google for 'bell' produces (can you guess?) a multitude of pictures of the type of bell that makes a pleasant ringing sound when one calls for afternoon tea, not to mention the more coarse cousins to be found in ships and belfreys.

A similar search on Live's search engine produces (can you guess?) an array of images of people with the name Bell.

Ask follows Google's bias whereas Yahoo does mix object and people, though the pictures of people are not, shall we say, quite what you might expect.

What is interesting about these differences is that there are obvious biases in the search engines in terms of how they disambiguate terms (is it an object or a person that the user intends to find?). However, none of the search engines really acknowledges the ambiguity, nor do they assist the user in dealing with it. Perhaps the best attempt here is in Ask's interactive query completion, however this is far from perfect as typing 'bell' precludes any suggestions that start with a term other than 'bell'. Thus someone with the surname 'bell' is not going to surface at this point in the interaction.

 

Comments

[Udi]

http://flickr.com/photos/tags/bell/clusters/

[Vitorino]

This is easy to solve: mix the results people, objects and other categories and then when the user pick an image she will indicate which category she wants to find pictures.

[Tony]

Let me take a different view. What if I don't know what Bell I am looking for? (Bell is probably a bad example) The most popular results will probably help me to determine which Bell I really wanted. Therefore, the "bias" helped me to narrow down my search.

Without the search engine asking you questions to narrow down the results you are going to get a wide range of items that are returned. However, If they start asking you questions, then people would probably go elsewhere to search quicker.
There is the advanced search feature where you can eliminate words or include specific phrases.

FYI - type in 'search engine bias' and you will get nearly 2 million results. Not having the time to weed through all of them, the most popular probably make up 99% of all the intended searches (just my guess, not proven or supported).

[random]

i think google's is better. if i'm actually looking for the picture of a bell, what am i supposed to type into yahoo to get such a regular bell?

i'm sure if i type the name of whoever that guy is (including his first name) into google i'll get the picture.

[Dmitry]

It doesn't disambiguate. In fact most search engines (Live and Yahoo for sure) have no understanding of human language at all. What you see is a result of (partially) supervised learning where labeled samples come from humans either through direct evaluation or through click analysis.

Google's results may also use their "game" where you label images for them for free. Here's the link: http://images.google.com/imagelabeler/. This may be why it recognizes a bell as a bell, rather than a person's name. I, for one, had no idea what the names of those folks were.

[Matthew Hurst]

Dmitry,

I suspect that some form of non trivial disambiguation does occur in some cases. Look at the results on Google for 'Michael Jordan' or 'William Cohen'. Both get a couple of different people on the first page even though we might assume quite a large difference between the individuals in terms of in-links or other measures of visibility. Live does show both William Cohens as does Yahoo. I can still believe, however, that this may be due to structural queues.

[hoon]

Why would I want a bunch of people when I searched for bell? If I typed Bell, that might trigger people over objects but really it seems like Live is a mistake doesn't it? (and Tinkerbell? WTF?)

[hoon]

Why would I want a bunch of people when I searched for bell? If I typed Bell, that might trigger people over objects but really it seems like Live is a mistake doesn't it? (and Tinkerbell? WTF?)