via Cornell University
When you search for something on the internet, do you scroll through page after page of suggestions – or pick from the first few choices?
Because most people choose from the tops of these lists, they rarely see the vast majority of the options, creating a potential for bias in everything from hiring to media exposure to e-commerce.
In a new paper, Cornell researchers introduce a tool they’ve developed to improve the fairness of online rankings without sacrificing their usefulness or relevance.
“If you could examine all your choices equally and then decide what to pick, that may be considered ideal. But since we can’t do that, rankings become a crucial interface to navigate these choices,” said computer science doctoral student Ashudeep Singh, co-first author of “Controlling Fairness and Bias in Dynamic Learning-to-Rank,” which won the Best Paper Award at the Association for Computing Machinery SIGIR Conference on Research and Development in Information Retrieval, held virtually July 25-30.
“For example, many YouTubers will post videos of the same recipe, but some of them get seen way more than others, even though they might be very similar,” Singh said. “And this happens because of the way search results are presented to us. We generally go down the ranking linearly and our attention drops off fast.”
The researchers’ method, called FairCo, gives roughly equal exposure to equally relevant choices and avoids preferential treatment for items that are already high on the list. This can correct the unfairness inherent in existing algorithms, which can exacerbate inequality and political polarization, and curtail personal choice.
“What ranking systems do is they allocate exposure. So how do we make sure that everybody receives their fair share of exposure?” said Thorsten Joachims, professor of computer science and information science, and the paper’s senior author. “What constitutes fairness is probably very different in, say, an e-commerce system and a system that ranks resumes for a job opening. We came up with computational tools that let you specify fairness criteria, as well as the algorithm that will provably enforce them.”
Online ranking systems were originally based on library science from the 1960s and ’70s, which sought to make it easier for users to find the books they wanted. But this approach can be unfair in two-sided markets, in which one entity wants to find something and another wants to be found.
“Much of machine learning work in optimizing rankings is still very much focused on maximizing utility to the users,” Joachims said. “What we’ve done over the last few years is come up with notions of how to maximize utility while still being fair to the items that are being searched.”
Algorithms that prioritize more popular items can be unfair because the higher a choice appears in the list, the more likely users are to click on and react to it. This creates a “rich get richer” phenomenon where one choice becomes increasingly popular, and other choices go unseen.
Algorithms also seek the most relevant items to searchers, but because the vast majority of people choose one of the first few items in a list, small differences in relevance can lead to huge discrepancies in exposure. For example, if 51% of the readers of a news publication prefer opinion pieces that skew conservative, and 49% prefer essays that are more liberal, all of the top stories highlighted on the home page could conceivably lean conservative, according to the paper.
“When small differences in relevance lead to one side being amplified, that often causes polarization, where some people tend to dominate the conversation and other opinions get dropped without their fair share of attention,” Joachims said. “You might want to use it in an e-commerce system to make sure that if you’re producing a product that 30% of people like, you’re getting a certain amount of exposure based on that. Or if you have a resume database, you could formulate safeguards to make sure it’s not discriminating by race or gender.”
The Latest Updates from Bing News & Google News
Go deeper with Bing News on:
Online ranking systems
- American Public University System Named Most Popular GI Bill School by Military Timeson October 5, 2020 at 5:51 am
CHARLES TOWN, W.Va., Oct. 5, 2020 /PRNewswire/ -- (APUS), which offers over 200 online degree and certificate programs through American Military University (AMU) and American Public ...
- Ranked choice voting: a failed experimenton October 4, 2020 at 6:00 am
Ballot Measure 2 proposes to change Alaska’s election laws by asking Alaska voters to make ranked choice voting (RCV) the new election law in Alaska. Ranked choice voting (also known ...
- WatcH Abilene Christian vs Army Live: College football 2020 Online TVon October 3, 2020 at 9:36 am
Abilene Christian vs Army Live, The last time Florida rose to No. 3 in the AP Top 25 rankings, its opponent in its next game was South Carolina and the Gators throttled the Gamecocks 44-11 in 2012.
- Watch South Carolina vs Florida Live Stream Reddit Online NCAA Week 5on October 3, 2020 at 8:37 am
This I Think Watch South Carolina vs Florida Live,The last time Florida rose to No. 3 in the AP Top 25 rankings, its opponent in its next game was South Carolina and the Gators throttled the Gamecocks ...
- The FDA Just Ranked Coronavirus Tests by Sensitivity. Here's What That Meanson October 3, 2020 at 4:11 am
Diagnostic tests for the coronavirus come in all shapes and sizes, and they vary dramatically in quality as well. One key test metric, sensitivity, is particularly important because it determines the ...
Go deeper with Google Headlines on:
Online ranking systems
Go deeper with Bing News on:
Fairer search results
- Naver Fined $23 Million for Manipulating Search Algorithmson October 6, 2020 at 4:33 pm
Korea's top search engine Naver has been issued a fine of almost W27 billion for manipulating search results to block its competitors and mislead consumers (US$1=W1,162). The fine was handed down ...
- Naver fined for search manipulationon October 6, 2020 at 4:26 pm
The South Korean search giant was fined 26.7 billion won, around $23 million, from the South Korean FTC for manipulating its search algorithm to show rivals lower.
- South Korean Search Giant Fined $23 Mln For Manipulating Resultson October 6, 2020 at 4:16 pm
South Korea's biggest internet portal Naver was hit with an unprecedented 26.7 billion won ($23 million) fine Tuesday for manipulating search results in favour of its own online shopping service.
- S Korea’s Naver fined US$23m for manipulating resultson October 6, 2020 at 9:31 am
South Korea’s biggest Internet portal, Naver Corp, was yesterday hit with an unprecedented 26.7 billion won (US$23 million) fine for manipulating search results in favor of its own online shopping ...
- South Korean search giant fined $23 million for manipulating resultson October 6, 2020 at 3:55 am
South Korea's biggest internet portal Naver was hit with an unprecedented 26.7 billion won ($23 million) fine Tuesday for manipulating search results in favor of its own online shopping service.