A mystery in the machine

Professor in Department of Computer Science and Engineering, University of Washington‌‌

Charlotte Moreau

Algorithms lie at the heart of machine learning, which, in turn lies at the heart of much of modern life–from online shopping to intelligence gathering. But most of us know little about these powerful tools and how they work. Is this wise?

You may not know it, but machine learning is all around you. When you type a query into a search engine, it’s how the engine figures out which results to show you–and which ads. When you read your email, you don’t see most of the spam because machine learning has filtered it out. Go to Amazon to buy a book or Netflix to watch a video, and a machine-learning system helpfully recommends some others you might like. Facebook uses machine learning to decide which updates to show you, and Twitter does the same for tweets. Whenever you use a computer, chances are machine learning is involved somewhere.

Traditionally, the only way to get a computer to do something– from adding two numbers together to flying an aeroplane–was to write down an algorithm explaining how, in painstaking detail. But machine-learning algorithms are different: they figure it out on their own, by making inferences from data. And the more data they have, the better they get. Now we don’t have to programme computers; they programme themselves.

It’s not just in cyberspace, either. Machine learning plays a part in every stage of our lives: grading college applicants’ exams, screening job candidates, picking stocks, diagnosing illnesses, finding partners. Society is changing, one learning algorithm at a time. Machine learning is remaking science, technology, business, politics, and war. Satellites, DNA sequencers, and particle accelerators probe nature in ever finer detail, and learning algorithms turn the torrents of data into new scientific knowledge. Companies know their customers like never before, and political candidates know their supporters. Unmanned vehicles pilot themselves across land, sea and air. No one programmed your tastes into Amazon’s recommendation system–it used your purchasing history to figure them out on its own. Google’s self-driving car taught itself how to stay on the road; no engineer wrote an algorithm instructing it how to get from A to B.

Machine learning is something new under the sun–a technology that builds itself. Ever since our remote ancestors started sharpening stones into tools, it is only humans who have designed artefacts. But learning algorithms are artefacts that design other artefacts. A learning algorithm is like a master craftsman: every one of its productions is different, and exquisitely matched to the customer’s needs. But instead of turning stone into masonry or gold into jewellery, machine learning turns data into algorithms. And the more data it has, the more intricate the algorithms.

Humans are the species that adapts the world to itself instead of adapting itself to the world. Machine learning is the newest chapter in this million-year saga: with it, the world senses what you want and changes accordingly, without you having to lift a finger. Like a magic forest, your surroundings–virtual today, physical tomorrow–rearrange themselves as you move through them. The path you picked out between the trees and bushes grows into a road. Signs pointing the way spring up in the places you got lost.

These seemingly magical technologies work because, at its core, machine learning is about prediction–predicting what we want, the results of our actions, how to achieve our goals, how the world will change. Once upon a time, we relied on shamans and soothsayers for this, but they were much too fallible. Science’s predictions are more trustworthy, but they are limited to what we can systematically observe and tractably model. Big data and machine learning greatly expand that scope. Some everyday things can be predicted by the unaided mind, from catching a ball to carrying on a conversation. Some things, try as we might, are just unpredictable. For the vast middle ground between the two, there’s machine learning.

Paradoxically, even as they open new windows on nature and human behaviour, learning algorithms themselves have remained shrouded in mystery. Hardly a day goes by without a story in the media involving machine learning, whether it’s a tech company launching a virtual personal assistant; Google’s AlphaGo beating the human Go champion; US retailer Target finding out a teenager is pregnant before her parents do; or the US National Security Agency (NSA) looking for dots to connect. But in each case the learning algorithm driving the story is a black box. Even books on big data skirt around what really happens when the computer swallows all those terabytes and magically comes up with new insights. At best, we’re left with the impression that learning algorithms just find correlations between pairs of events, such as Googling “flu medicine” and having the flu. But finding correlations is to machine learning no more than bricks are to houses, and people don’t live in bricks.

When a new technology is as pervasive and game-changing as machine learning, it’s not wise to let it remain a black box. Opacity opens the door to error and misuse. Amazon’s algorithm, more than any one person, determines what books are read in the world today. The NSA’s algorithms decide who is, or isn’t, a potential terrorist. Climate models decide what’s a safe level of carbon dioxide in the atmosphere. Stock-picking models drive the economy more than most of us do. You can’t control what you don’t understand, and that’s why you need to understand machine learning–as a citizen, a professional, and a human being engaged in the pursuit of happiness.

Domingos, Pedro (2015), The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World, Basic Books

See http://homes.cs.washington.edu/~pedrod/

Innovation and the digital economy at the OECD Forum 2016

The future of work at the OECD Forum 2016

Other OECD Forum 2016 issues

OECD work on science and technology

OECD work on the Internet economy

OECD work on innovation

© OECD Yearbook 2016




Economic data

E-Newsletter

Stay up-to-date with the latest news from the OECD by signing up for our e-newsletter :

Twitter feed

Suscribe now

<b>Subscribe now!</b>

To receive your exclusive print editions delivered to you directly


Online edition
Previous editions

Don't miss

  • “Nizip” refugee camp visit
    July 2016: OECD Secretary-General Angel Gurría visits the “Nizip” refugee camp, situated between Gaziantep and the Turkish-Syrian border, accompanied by Turkey’s Deputy Prime Minister Mehmet Şimşek. The camp accommodates a small number of the 2.75 million Syrians currently registered in Turkey, mostly outside the camps. In his tour of the camp, Mr Gurría visits a school, speaks with refugees and gives a short interview.
  • OECD Observer i-Sheet Series: OECD Observer i-Sheets are smart contents pages on major issues and events. Use them to find current or recent articles, video, books and working papers. To browse on paper and read on line, or simply download.
  • Queen Maxima of the Netherlands gives a speech next to Mexico's President Enrique Pena Nieto (not pictured) during the International Forum of Financial Inclusion at the National Palace in Mexico City, Mexico June 21, 2016.
  • How sustainable is the ocean as a source of economic development? The Ocean Economy in 2030 examines the risks and uncertainties surrounding the future development of ocean industries, the innovations required in science and technology to support their progress, their potential contribution to green growth and some of the implications for ocean management.
  • OECD Environment Director Simon Upton presented a talk at Imperial College London on 21 April 2016. With the world awash in surplus oil and prices languishing around US$40 per barrel, how can governments step up efforts to transform the world’s energy systems in line with the Paris Agreement?
  • Happy 10th birthday to Twitter. This 2008 OECD Observer interview with Henry Copeland said you’d do well.
  • The OECD Gender Initiative examines existing barriers to gender equality in education, employment, and entrepreneurship. The gender portal monitors the progress made by governments to promote gender equality in both OECD and non-OECD countries and provides good practices based on analytical tools and reliable data.
  • Once migrants reach Europe, countries face integration challenge: OECD's Thomas Liebig speaks to NPR's Audie Cornish.

  • Message from the International Space Station to COP21

  • COP21 Will Get Agreement With Teeth: OECD Secretary-General Angel Gurría on Bloomberg

  • The carbon clock is ticking: OECD’s Gurría on CNBC

  • If we want to reach zero net emissions by the end of the century, we must align our policies for a low-carbon economy, put a price on carbon everywhere, spend less subsidising fossil fuels and invest more in clean energy. OECD at #COP21 – OECD statement for #COP21
  • They are green and local --It’s a new generation of entrepreneurs in Kenya with big dreams of sustainable energy and the drive to see their innovative technologies throughout Africa. blogs.worldbank.org
  • Pole to Paris Project
  • In order to face global warming, Asia needs at least $40 billion per year, derived from both the public and private sector. Read how to bridge the climate financing gap on the Asian Bank of Development's website.
  • How can cities fight climate change?
    Discover projects in Denmark, Canada, Australia, Japan and Mexico.
  • Climate: What's changed, what hasn't, what we can do about it.
    Lecture by OECD Secretary-General Angel Gurría, hosted by the London School of Economics and Aviva Investors in association with ClimateWise, London, UK, 3 July 2015.

  • Climate change: “We should not disagree when scientists tell us we have a window of opportunity–10-15 years–to turn this thing around” argues Senator Bernie Sanders.

  • In the long-run, the EU benefits from migration, says OECD Head of International Migration Division Jean-Christophe Dumont.
  • Is technological progress slowing down? Is it speeding up? At the OECD, we believe the research from our Future of ‪Productivity‬ project helps to resolve this paradox.
  • Is inequality bad for growth? That redistribution boosts economies is not established by the evidence says FT economics editor Chris Giles. Read more on www.ft.com.
  • Catherine Mann, OECD Chief Economist, explains on Bloomberg why "too much bank lending can slow economic growth".
  • Interested in a career in Paris at the OECD? The OECD is a major international organisation, with a mission to build better policies for better lives. With our hub based in one of the world's global cities and offices across continents, find out more at www.oecd.org/careers .

Most Popular Articles

Poll

What issue are you most concerned about in 2016?

Unemployment
Euro crisis
International conflict
Global warming
Other

OECD Insights Blog

NOTE: All signed articles in the OECD Observer express the opinions of the authors
and do not necessarily represent the official views of OECD member countries.

All rights reserved. OECD 2016