Nav: Home

Approximating a kernel of truth

March 10, 2020

By using an approximate rather than explicit "kernel" function to extract relationships in very large data sets, KAUST researchers have been able to dramatically accelerate the speed of machine learning. The approach promises to greatly improve the speed of artificial intelligence (AI) in the era of big data.

When AI is exposed to a large unknown data set, it needs to analyze the data and develop a model or function that describes the relationships in the set. The calculation of this function, or kernel, is a computationally intensive task that increases in complexity cubically (to the power of three) with the size of the data set. In the era of big data and increasing reliance on AI for analysis, this presents a real problem where kernel selection can become impractically time consuming.

With the supervision of Xin Gao, Lizhong Ding and his colleagues have been working on methods to speed up kernel selection using statistics.

"The computational complexity of accurate kernel selection is usually cubic with the number of samples," says Ding. "This kind of cubic scaling is prohibitive for big data. We have instead proposed an approximation approach for kernel selection, which significantly improves the efficiency of kernel selection without sacrificing predictive performance."

The true or accurate kernel provides a verbatim description of relationships in the data set. What the researchers found is that statistics can be used to derive an approximate kernel that is almost as good as the accurate version, but can be computed many times faster, scaling linearly, rather than cubically, with the size of the data set.

To develop the approach, the team had to construct specifically designed kernel matrices, or mathematical arrays, that could be computed quickly. They also had to establish the rules and theoretical bounds for selection of the approximate kernel that would still guarantee learning performance.

"The main challenge was that we needed to design new algorithms satisfying these two points at the same time," says Ding.

Combining a process of error estimation and mathematical approximation, the researchers were able to prove that their approximate kernel remains consistent with the accurate kernel and then demonstrated its performance in real examples.

"We have shown that approximate methods, such as our computing framework, provide sufficient accuracy for solving a kernel-based learning method, without the impractical computational burden of accurate methods," says Ding. "This provides an effective and efficient solution for problems in data mining and bioinformatics that require scalability."
-end-


King Abdullah University of Science & Technology (KAUST)

Related Relationships Articles:

'Feeling obligated' can impact relationships during social distancing
In a time where many are practicing 'social distancing' from the outside world, people are relying on their immediate social circles more than usual.
We can make predictions about relationships - but is this necessary?
'Predictions as to the longevity of a relationship are definitely possible,' says Dr Christine Finn from the University of Jena.
Disruptions of salesperson-customer relationships. Is that always bad?
Implications from sales relationship disruptions are intricate and can be revitalizing.
Do open relationships really work?
Open relationships typically describe couples in which the partners have agreed on sexual activity with someone other than their primary romantic partner, while maintaining the couple bond.
The 7 types of sugar daddy relationships
University of Colorado Denver researcher looks inside 48 sugar daddy relationships to better understand the different types of dynamics, break down the typical stereotype(s) and better understand how these relationships work in the United States.
Positive relationships boost self-esteem, and vice versa
Does having close friends boost your self-esteem, or does having high self-esteem influence the quality of your friendships?
Strong family relationships may help with asthma outcomes for children
Positive family relationships might help youth to maintain good asthma management behaviors even in the face of difficult neighborhood conditions, according to a new Northwestern University study.
In romantic relationships, people do indeed have a 'type'
Researchers at the University of Toronto show that people do indeed have a 'type' when it comes to dating, and that despite best intentions to date outside that type -- for example, after a bad relationship -- some will gravitate to similar partners.
Advancing dementia and its effect on care home relationships
New research published today in the journal Dementia by researchers from the University of Chichester focuses on the effects of behavioral change due to dementia in a residential care home setting.
Passion trumps love for sex in relationships
When women distinguish between sex and the relational and emotional aspects of a relationship, this determines how often couples in long-term relationships have sex.
More Relationships News and Relationships Current Events

Trending Science News

Current Coronavirus (COVID-19) News

Top Science Podcasts

We have hand picked the top science podcasts of 2020.
Now Playing: TED Radio Hour

Climate Mindset
In the past few months, human beings have come together to fight a global threat. This hour, TED speakers explore how our response can be the catalyst to fight another global crisis: climate change. Guests include political strategist Tom Rivett-Carnac, diplomat Christiana Figueres, climate justice activist Xiye Bastida, and writer, illustrator, and artist Oliver Jeffers.
Now Playing: Science for the People

#562 Superbug to Bedside
By now we're all good and scared about antibiotic resistance, one of the many things coming to get us all. But there's good news, sort of. News antibiotics are coming out! How do they get tested? What does that kind of a trial look like and how does it happen? Host Bethany Brookeshire talks with Matt McCarthy, author of "Superbugs: The Race to Stop an Epidemic", about the ins and outs of testing a new antibiotic in the hospital.
Now Playing: Radiolab

Speedy Beet
There are few musical moments more well-worn than the first four notes of Beethoven's Fifth Symphony. But in this short, we find out that Beethoven might have made a last-ditch effort to keep his music from ever feeling familiar, to keep pushing his listeners to a kind of psychological limit. Big thanks to our Brooklyn Philharmonic musicians: Deborah Buck and Suzy Perelman on violin, Arash Amini on cello, and Ah Ling Neu on viola. And check out The First Four Notes, Matthew Guerrieri's book on Beethoven's Fifth. Support Radiolab today at Radiolab.org/donate.