Saturday, August 16, 2008

Thinking Outside the Black Box

I was reading Techmeme today, and I noticed an LA Times article about RushmoreDrive, described on its About Us page as "a first-of-its-kind search engine for the Black community." My first reaction, blogged by others already, was that this idea was dumb and racist. In fact, it took some work to find positive commentary about RushmoreDrive.

But I've learned from the way the blogosphere handled the Cuil launch not to trust anyone who evaluates a search engine without having tried it, myself included. My wife and I have been the only white people at Amy Ruth's and the service was as gracious as the chicken and waffles were delicious; I decided I'd try my luck on a search engine not targeted at my racial profile.

The search quality is solid, comparable to that of Google, Yahoo, and Microsoft. In fact, the site looks a lot like a re-skinning (no pun intended) of Ask.com, a corporate sibling of IAC-owned RushmoreDrive. Like Ask.com, RushmoreDrive emphasizes search refinement through narrowing and broadening refinements.

What I find ironic is that the whole controversy about racial bias in relevance ranking reveals the much bigger problem--that relevance ranking should not be a black box (ok, maybe this time I'll take responsibility for the pun). I've been beating this drum at The Noisy Channel ever since I criticized Amit Singhal for Google's lack of transparency. I think that sites like RushmoreDrive are inevitable if search engines refuse to cede more control of search results to users.

I don't know how much information race provides as prior to influence statistical ranking approaches, but I'm skeptical that the effects are useful or even noticeable beyond a few well-chosen examples. I'm more inclined to see RushmoreDrive as a marketing ploy by the folks at IAC--and perhaps a successful one. I doubt that Google is running scared, but I think this should be a wake-up call to folks who are convinced that personalized relevance ranking is the end goal of user experience for search engines.

7 comments:

Anonymous said...

First, its good to see you all give this thing a try instead of making silly suggestions that it is racist.

I have used RushmoreDrive.com and find it much better than Google, but I'm black so that's no surprise -- and thats the point of it, right? It's not perfect -- but its better at finding things that relate to us. I like that the search results are not just links, but include video, images, and news. Also, i tested rushmores results against Ask.com and they are different so I don't think it's just a skinned Ask.com.

As a member of Black community, I do appreciate that you gave it a try.

Daniel Tunkelang said...

Intellectual honesty is color blind, so I try my best. I'm glad it's appreciated.

I'm actually giving RushmoreDrive an extended trial: I made it the default search for my browser. And you're right, the results are different from Ask.com--I can see that myself now after more testing.

But the differences aren't exactly black and white. On one hand, the query "apollo" (on RushmoreDrive vs. on Ask) performs as I'd expected: RushmoreDrive includes a results related to the Apollo Theater in its results, but Ask does not--through both include Apollo Theater in the Narrow Your Search list. On the other hand, the query "soca" (on RushmoreDrive vs. on Ask) surprised me: RushmoreDrive's first two results are the Serious Organised Crime Agency and the Soccer Organization of Charlottesville-Albemarle, while the top five results from Ask relate to soca music.

It strikes me as tough to tune a single relevance ranking algorithm to a population that includes over one in eight Americans. Granted, it's even tougher to tune to the entire population of the US (I'm assuming that the global web search engines tune by country). I think that, fundamentally, a search engine shouldn't be trying to stereotype me as white, black, or anything else, but rather should be helping me articulate my unique personal search goals. And everyday I'm less convinced that tweaking the relevance ranking function is the way to get there.

In any case, I do think the similarity in presentation style between RushmoreDrive and Ask is more than a coincidence, I found this blog post from April: IAC to Launch a Flurry of Niche Sites. I don't know how much IAC is able to reuse its technology and content assets to create these sites, but it strikes me as a clever business strategy.

Daniel Tunkelang said...

Coverage by Sam Diaz of ZDNet: Searching by color to nibble away at Google

L. Wu said...

Another way to look at it would be to consider a different definition of racism:

http://theangryblackwoman.wordpress.com/2006/05/07/some-things-you-need-to-understand-1/

I think it is problematic if one search engine knowingly or unknowingly wields power and/or unconscious prejudice, and lets that seep into how folks get directed to different sites.

Search results aside, Google has not seemed very responsive to complaints about their sponsored ads, yet another arena in which Google rather than its competitors could be argued to support institutional forms of various -isms.

~L

Daniel Tunkelang said...

Google does not disclose exactly how it ranks results, but I think it's fair to assert that they strive to be "democratic" by treating a hyperlinks as a voting scheme to determine the query-independent authority of a page. To best of my knowledge, all major search engines, including RushmoreDrive, take this approach.

I'm not sure how you would define, let alone determine, whether a search engine unknowingly wields power or unconscious prejudice.

For example, Google and Yahoo often return Wikipedia entries as top results. Cuil does not. Are Google and Yahoo prejudiced in favor of Wikipedia? Is Cuil prejudiced against it? And, in either case, what would be the broader implications of such a prejudice?

I return to my main point: search engines should not be making value judgments; they should be helping people make their own value judgments.

Anonymous said...

Why dont people type what thye really want to find, in the article in the times he says that a search of whitney brought up a bank and not whitney houston WHY DONT YOU TYPE HER FULL NAME AND YOU CAN GET A FULL PAGE OF CRAP ABOUT HER. Its a dumb idea overall, just another way to seperate people of different races.

Daniel Tunkelang said...

No need to shout. And let's face it, these one-word examples are all pretty artificial. It would be interesting to see if / how the search engines differ on a more realistic queries.

But I maintain my previous point: black-box relevance ranking is inherently a guessing game, at least for queries where it isn't possible to determine the user's intent reliably from the query. In principle, knowing more about the user should better inform that guess. e.g., by improving the document priors. But I think that guessing the user's intent is a wrong-headed approach when, as Feynman famously said, you can just ask.

Saturday, August 16, 2008

Thinking Outside the Black Box

I was reading Techmeme today, and I noticed an LA Times article about RushmoreDrive, described on its About Us page as "a first-of-its-kind search engine for the Black community." My first reaction, blogged by others already, was that this idea was dumb and racist. In fact, it took some work to find positive commentary about RushmoreDrive.

But I've learned from the way the blogosphere handled the Cuil launch not to trust anyone who evaluates a search engine without having tried it, myself included. My wife and I have been the only white people at Amy Ruth's and the service was as gracious as the chicken and waffles were delicious; I decided I'd try my luck on a search engine not targeted at my racial profile.

The search quality is solid, comparable to that of Google, Yahoo, and Microsoft. In fact, the site looks a lot like a re-skinning (no pun intended) of Ask.com, a corporate sibling of IAC-owned RushmoreDrive. Like Ask.com, RushmoreDrive emphasizes search refinement through narrowing and broadening refinements.

What I find ironic is that the whole controversy about racial bias in relevance ranking reveals the much bigger problem--that relevance ranking should not be a black box (ok, maybe this time I'll take responsibility for the pun). I've been beating this drum at The Noisy Channel ever since I criticized Amit Singhal for Google's lack of transparency. I think that sites like RushmoreDrive are inevitable if search engines refuse to cede more control of search results to users.

I don't know how much information race provides as prior to influence statistical ranking approaches, but I'm skeptical that the effects are useful or even noticeable beyond a few well-chosen examples. I'm more inclined to see RushmoreDrive as a marketing ploy by the folks at IAC--and perhaps a successful one. I doubt that Google is running scared, but I think this should be a wake-up call to folks who are convinced that personalized relevance ranking is the end goal of user experience for search engines.

7 comments:

Anonymous said...

First, its good to see you all give this thing a try instead of making silly suggestions that it is racist.

I have used RushmoreDrive.com and find it much better than Google, but I'm black so that's no surprise -- and thats the point of it, right? It's not perfect -- but its better at finding things that relate to us. I like that the search results are not just links, but include video, images, and news. Also, i tested rushmores results against Ask.com and they are different so I don't think it's just a skinned Ask.com.

As a member of Black community, I do appreciate that you gave it a try.

Daniel Tunkelang said...

Intellectual honesty is color blind, so I try my best. I'm glad it's appreciated.

I'm actually giving RushmoreDrive an extended trial: I made it the default search for my browser. And you're right, the results are different from Ask.com--I can see that myself now after more testing.

But the differences aren't exactly black and white. On one hand, the query "apollo" (on RushmoreDrive vs. on Ask) performs as I'd expected: RushmoreDrive includes a results related to the Apollo Theater in its results, but Ask does not--through both include Apollo Theater in the Narrow Your Search list. On the other hand, the query "soca" (on RushmoreDrive vs. on Ask) surprised me: RushmoreDrive's first two results are the Serious Organised Crime Agency and the Soccer Organization of Charlottesville-Albemarle, while the top five results from Ask relate to soca music.

It strikes me as tough to tune a single relevance ranking algorithm to a population that includes over one in eight Americans. Granted, it's even tougher to tune to the entire population of the US (I'm assuming that the global web search engines tune by country). I think that, fundamentally, a search engine shouldn't be trying to stereotype me as white, black, or anything else, but rather should be helping me articulate my unique personal search goals. And everyday I'm less convinced that tweaking the relevance ranking function is the way to get there.

In any case, I do think the similarity in presentation style between RushmoreDrive and Ask is more than a coincidence, I found this blog post from April: IAC to Launch a Flurry of Niche Sites. I don't know how much IAC is able to reuse its technology and content assets to create these sites, but it strikes me as a clever business strategy.

Daniel Tunkelang said...

Coverage by Sam Diaz of ZDNet: Searching by color to nibble away at Google

L. Wu said...

Another way to look at it would be to consider a different definition of racism:

http://theangryblackwoman.wordpress.com/2006/05/07/some-things-you-need-to-understand-1/

I think it is problematic if one search engine knowingly or unknowingly wields power and/or unconscious prejudice, and lets that seep into how folks get directed to different sites.

Search results aside, Google has not seemed very responsive to complaints about their sponsored ads, yet another arena in which Google rather than its competitors could be argued to support institutional forms of various -isms.

~L

Daniel Tunkelang said...

Google does not disclose exactly how it ranks results, but I think it's fair to assert that they strive to be "democratic" by treating a hyperlinks as a voting scheme to determine the query-independent authority of a page. To best of my knowledge, all major search engines, including RushmoreDrive, take this approach.

I'm not sure how you would define, let alone determine, whether a search engine unknowingly wields power or unconscious prejudice.

For example, Google and Yahoo often return Wikipedia entries as top results. Cuil does not. Are Google and Yahoo prejudiced in favor of Wikipedia? Is Cuil prejudiced against it? And, in either case, what would be the broader implications of such a prejudice?

I return to my main point: search engines should not be making value judgments; they should be helping people make their own value judgments.

Anonymous said...

Why dont people type what thye really want to find, in the article in the times he says that a search of whitney brought up a bank and not whitney houston WHY DONT YOU TYPE HER FULL NAME AND YOU CAN GET A FULL PAGE OF CRAP ABOUT HER. Its a dumb idea overall, just another way to seperate people of different races.

Daniel Tunkelang said...

No need to shout. And let's face it, these one-word examples are all pretty artificial. It would be interesting to see if / how the search engines differ on a more realistic queries.

But I maintain my previous point: black-box relevance ranking is inherently a guessing game, at least for queries where it isn't possible to determine the user's intent reliably from the query. In principle, knowing more about the user should better inform that guess. e.g., by improving the document priors. But I think that guessing the user's intent is a wrong-headed approach when, as Feynman famously said, you can just ask.

Saturday, August 16, 2008

Thinking Outside the Black Box

I was reading Techmeme today, and I noticed an LA Times article about RushmoreDrive, described on its About Us page as "a first-of-its-kind search engine for the Black community." My first reaction, blogged by others already, was that this idea was dumb and racist. In fact, it took some work to find positive commentary about RushmoreDrive.

But I've learned from the way the blogosphere handled the Cuil launch not to trust anyone who evaluates a search engine without having tried it, myself included. My wife and I have been the only white people at Amy Ruth's and the service was as gracious as the chicken and waffles were delicious; I decided I'd try my luck on a search engine not targeted at my racial profile.

The search quality is solid, comparable to that of Google, Yahoo, and Microsoft. In fact, the site looks a lot like a re-skinning (no pun intended) of Ask.com, a corporate sibling of IAC-owned RushmoreDrive. Like Ask.com, RushmoreDrive emphasizes search refinement through narrowing and broadening refinements.

What I find ironic is that the whole controversy about racial bias in relevance ranking reveals the much bigger problem--that relevance ranking should not be a black box (ok, maybe this time I'll take responsibility for the pun). I've been beating this drum at The Noisy Channel ever since I criticized Amit Singhal for Google's lack of transparency. I think that sites like RushmoreDrive are inevitable if search engines refuse to cede more control of search results to users.

I don't know how much information race provides as prior to influence statistical ranking approaches, but I'm skeptical that the effects are useful or even noticeable beyond a few well-chosen examples. I'm more inclined to see RushmoreDrive as a marketing ploy by the folks at IAC--and perhaps a successful one. I doubt that Google is running scared, but I think this should be a wake-up call to folks who are convinced that personalized relevance ranking is the end goal of user experience for search engines.

7 comments:

Anonymous said...

First, its good to see you all give this thing a try instead of making silly suggestions that it is racist.

I have used RushmoreDrive.com and find it much better than Google, but I'm black so that's no surprise -- and thats the point of it, right? It's not perfect -- but its better at finding things that relate to us. I like that the search results are not just links, but include video, images, and news. Also, i tested rushmores results against Ask.com and they are different so I don't think it's just a skinned Ask.com.

As a member of Black community, I do appreciate that you gave it a try.

Daniel Tunkelang said...

Intellectual honesty is color blind, so I try my best. I'm glad it's appreciated.

I'm actually giving RushmoreDrive an extended trial: I made it the default search for my browser. And you're right, the results are different from Ask.com--I can see that myself now after more testing.

But the differences aren't exactly black and white. On one hand, the query "apollo" (on RushmoreDrive vs. on Ask) performs as I'd expected: RushmoreDrive includes a results related to the Apollo Theater in its results, but Ask does not--through both include Apollo Theater in the Narrow Your Search list. On the other hand, the query "soca" (on RushmoreDrive vs. on Ask) surprised me: RushmoreDrive's first two results are the Serious Organised Crime Agency and the Soccer Organization of Charlottesville-Albemarle, while the top five results from Ask relate to soca music.

It strikes me as tough to tune a single relevance ranking algorithm to a population that includes over one in eight Americans. Granted, it's even tougher to tune to the entire population of the US (I'm assuming that the global web search engines tune by country). I think that, fundamentally, a search engine shouldn't be trying to stereotype me as white, black, or anything else, but rather should be helping me articulate my unique personal search goals. And everyday I'm less convinced that tweaking the relevance ranking function is the way to get there.

In any case, I do think the similarity in presentation style between RushmoreDrive and Ask is more than a coincidence, I found this blog post from April: IAC to Launch a Flurry of Niche Sites. I don't know how much IAC is able to reuse its technology and content assets to create these sites, but it strikes me as a clever business strategy.

Daniel Tunkelang said...

Coverage by Sam Diaz of ZDNet: Searching by color to nibble away at Google

L. Wu said...

Another way to look at it would be to consider a different definition of racism:

http://theangryblackwoman.wordpress.com/2006/05/07/some-things-you-need-to-understand-1/

I think it is problematic if one search engine knowingly or unknowingly wields power and/or unconscious prejudice, and lets that seep into how folks get directed to different sites.

Search results aside, Google has not seemed very responsive to complaints about their sponsored ads, yet another arena in which Google rather than its competitors could be argued to support institutional forms of various -isms.

~L

Daniel Tunkelang said...

Google does not disclose exactly how it ranks results, but I think it's fair to assert that they strive to be "democratic" by treating a hyperlinks as a voting scheme to determine the query-independent authority of a page. To best of my knowledge, all major search engines, including RushmoreDrive, take this approach.

I'm not sure how you would define, let alone determine, whether a search engine unknowingly wields power or unconscious prejudice.

For example, Google and Yahoo often return Wikipedia entries as top results. Cuil does not. Are Google and Yahoo prejudiced in favor of Wikipedia? Is Cuil prejudiced against it? And, in either case, what would be the broader implications of such a prejudice?

I return to my main point: search engines should not be making value judgments; they should be helping people make their own value judgments.

Anonymous said...

Why dont people type what thye really want to find, in the article in the times he says that a search of whitney brought up a bank and not whitney houston WHY DONT YOU TYPE HER FULL NAME AND YOU CAN GET A FULL PAGE OF CRAP ABOUT HER. Its a dumb idea overall, just another way to seperate people of different races.

Daniel Tunkelang said...

No need to shout. And let's face it, these one-word examples are all pretty artificial. It would be interesting to see if / how the search engines differ on a more realistic queries.

But I maintain my previous point: black-box relevance ranking is inherently a guessing game, at least for queries where it isn't possible to determine the user's intent reliably from the query. In principle, knowing more about the user should better inform that guess. e.g., by improving the document priors. But I think that guessing the user's intent is a wrong-headed approach when, as Feynman famously said, you can just ask.