Science Relies On Computer Modelling – So What Happens When It Goes Wrong?

Science Relies On Computer Modelling – So What Happens When It Goes Wrong?

From the transforming discovery of penicillin to the theories of relativity and quantum mechanics, science progressed with mind-boggling speed even before there were computers. Much of this is down to the robustness of the scientific method: scientific results are validated by being replicated and extended by other scientists.

But the way we do science is changing – we now rely increasingly on complex computer models to understand nature. And it turns out that these models can be nearly impossible to reproduce – meaning an important touchstone of science is being challenged. So what are the real-world repercussions of this change and what can we do about it?

Pre-modern science – known as “natural philosophy” – was empirical. Empirical science uses past observations to make predictions about the future, which may then be tested. Tycho Brahe, a 16th-century Danish astronomer, managed to make accurate and comprehensive observations of the heavens in this way.

Modern science, however, is theoretical. Theoretical science also makes predictions, but it derives them from mathematical models rather than from prior observations. Think of Isaac Newton’s laws of motion, such as the inverse square law of gravitation.

For example, there is an equation describing the orbit of the Earth around the sun. This equation can be used to build a computer model into which you can just plug certain variables and see how the solution changes. You could just plug in a future date and read off the position of the Earth at that date. You could also use the same program to model other planetary systems – it’s all based on the same mathematics. All you have to do is plug in different masses and various other properties of the bodies involved.

Such mathematical equations are great when they are available – but often they are not. For example, we know that there is no simple equation that solves the so-called “three-body problem”, which describes three bodies orbiting around and influencing each other by gravitational forces – like the moon, Earth and sun.

Much of current science deals with even more complicated systems, and similarly lacks exact solutions. Such models have to be “computational” – describing how a system changes from one instant to the next. But there is no way to determine the exact state at some time in the future other than by “simulating” its evolution in this way. Weather forecasting is a familiar example; until the advent of computers in the 1950s, it was impossible to predict future weather faster than it actually happened.

Current science typically consists of devising a mathematical model that describes a complicated system, then turning this into a computational simulation, and running the simulation to make predictions in order to validate the model.

When modelling fails

Modelling is used across scientific fields – ranging from astrophysics and climate prediction to bioinformatics and economics. But there is increasing debate about the fact that this science is difficult to validate through reproduction.

It turns out that simply describing experimental methods in words is not enough. That’s partly because natural languages such as English are simply too vague for describing computations precisely. There is, after all, a reason why programmers use programming languages. One of the biggest challenges in software development is in converting vague requirements into precise specifications of behaviour.

Humans – even scientists – are after all fallible. Transforming any information into a programme almost invariably introduces bugs along the way. For example, many scientists depend on data exploration tools such as spreadsheets, which are designed for ease of use and not for robustness. It is very easy simply to sum up the wrong range of cells in a spreadsheet, without getting any warnings. This was one of the methodological flaws in a paper that the US Republican Party used to base their pro-austerity policies on.

Similarly a recent study on the 15,770 spreadsheets that were made public during the investigation into the US corporation Enron showed that 24% of the spreadsheets containing at least one formula had obvious bugs, such as adding up blank cells.

In the natural sciences, the Mars Climate Observer, a space probe launched in 1998 to study the climate on Mars, was lost a year later because one part of the control software mistakenly used imperial instead of metric units. Another study of nine independent implementations of the same geoscience experiment – using the same dataset, algorithms, and programming language – showed very little agreement in the results obtained.

What’s more, even if the reader of a research paper can successfully interpret the writer’s precise meaning, and then faultlessly translate it into a program, there are still pitfalls in executing it. One particularly tricky class of problems arises from how computers handle numbers: although they can manipulate integers such as 42 and -17 with perfect accuracy, standard techniques for manipulating real numbers such π≈3.14 and √2≈1.414 permit only approximate accuracy. These approximations mean that apparently equivalent ways of computing the same value can yield different results.

So, what can be done? If even expert software developers cannot reliably produce correct software, what hope is there for amateur programmers like scientists?

One line of work is to produce tools for designing “domain-specific” programming languages, each tailored to a particular class of problem, such as the behaviour of agents in economic markets or the diffusion of drugs across cells. These aim to make it much easier for specialists to describe computations directly in familiar terms, rather than having to encode them indirectly in a general-purpose programming language.

A second approach seeks to design more expressive but still user friendly “type systems” for programs. These would make it easier to catch “silly” errors, such as blank cells in spreadsheets, or mixing up values in different units. It cannot rule out all logic errors though. A third line is to develop usable libraries of code for exact arithmetic, avoiding the problems of approximation.

There is every chance these approaches can help fix the problem going forward, or at least eliminate some of the risk. After all, the world needs science and scientists need computers – that’s not likely to change anytime soon.

About The Author

Jeremy Gibbons, Professor of Computing, University of Oxford. Head of the Programming Languages research theme.

This article originally appeared on The Conversation

Related Book:

InnerSelf Market

Amazon

enafarzh-CNzh-TWdanltlfifrdeiwhihuiditjakomsnofaplptruesswsvthtrukurvi

follow InnerSelf on

facebook icontwitter iconyoutube iconinstagram iconpintrest iconrss icon

 Get The Latest By Email

Weekly Magazine Daily Inspiration

EVIDENCE

Blue water surrounded by dead white grass
Map tracks 30 years of extreme snowmelt across US
by Mikayla Mace-Arizona
A new map of extreme snowmelt events over the last 30 years clarifies the processes that drive rapid melting.
White sea ice in blue water with the sun setting reflected in the water
Earth’s frozen areas are shrinking 33K square miles a year
by Texas A&M University
The Earth’s cryosphere is shrinking by 33,000 square miles (87,000 square kilometers) per year.
wind turbines
A controversial US book is feeding climate denial in Australia. Its central claim is true, yet irrelevant
by Ian Lowe, Emeritus Professor, School of Science, Griffith University
My heart sank last week to see conservative Australian commentator Alan Jones championing a contentious book about…
image
Reuters' Hot List of climate scientists is geographically skewed: why this matters
by Nina Hunter, Post-Doctoral Researcher, University of KwaZulu-Natal
The Reuters Hot List of “the world’s top climate scientists” is causing a buzz in the climate change community. Reuters…
A person holds a shell in their hand in blue water
Ancient shells hint past high CO2 levels could return
by Leslie Lee-Texas A&M
Using two methods to analyze tiny organisms found in sediment cores from the deep seafloor, researchers have estimated…
image
Matt Canavan suggested the cold snap means global warming isn't real. We bust this and 2 other climate myths
by Nerilie Abram, Professor; ARC Future Fellow; Chief Investigator for the ARC Centre of Excellence for Climate Extremes; Deputy Director for the Australian Centre for Excellence in Antarctic Science, Australian National University
Senator Matt Canavan sent many eyeballs rolling yesterday when he tweeted photos of snowy scenes in regional New South…
Ecosystem sentinels sound alarm for the oceans
by Tim Radford
Sea birds are known as ecosystem sentinels, warning of marine loss. As their numbers fall, so could the riches of the…
Why Sea Otters Are Climate Warriors
Why Sea Otters Are Climate Warriors
by Zak Smith
In addition to being one of the cutest animals on the planet, sea otters help maintain healthy, carbon-absorbing kelp…

LATEST VIDEOS

The Great Climate Migration Has Begun
The Great Climate Migration Has Begun
by Super User
The climate crisis is forcing thousands around the world to flee as their homes become increasingly uninhabitable.
The Last Ice Age Tells Us Why We Need To Care About A 2℃ Change In Temperature
The Last Ice Age Tells Us Why We Need To Care About A 2℃ Change In Temperature
by Alan N Williams, et al
The latest report from the Intergovernmental Panel on Climate Change (IPCC) states that without a substantial decrease…
Earth Has Stayed Habitable For Billions Of Years – Exactly How Lucky Did We Get?
Earth Has Stayed Habitable For Billions Of Years – Exactly How Lucky Did We Get?
by Toby Tyrrell
It took evolution 3 or 4 billion years to produce Homo sapiens. If the climate had completely failed just once in that…
How Mapping The Weather 12,000 Years Ago Can Help Predict Future Climate Change
How Mapping The Weather 12,000 Years Ago Can Help Predict Future Climate Change
by Brice Rea
The end of the last ice age, around 12,000 years ago, was characterised by a final cold phase called the Younger Dryas.…
The Caspian Sea Is Set To Fall By 9 Metres Or More This Century
The Caspian Sea Is Set To Fall By 9 Metres Or More This Century
by Frank Wesselingh and Matteo Lattuada
Imagine you are on the coast, looking out to sea. In front of you lies 100 metres of barren sand that looks like a…
Venus Was Once More Earth-like, But Climate Change Made It Uninhabitable
Venus Was Once More Earth-like, But Climate Change Made It Uninhabitable
by Richard Ernst
We can learn a lot about climate change from Venus, our sister planet. Venus currently has a surface temperature of…
Five Climate Disbeliefs: A Crash Course In Climate Misinformation
The Five Climate Disbeliefs: A Crash Course In Climate Misinformation
by John Cook
This video is a crash course in climate misinformation, summarizing the key arguments used to cast doubt on the reality…
The Arctic Hasn't Been This Warm For 3 Million Years and That Means Big Changes For The Planet
The Arctic Hasn't Been This Warm For 3 Million Years and That Means Big Changes For The Planet
by Julie Brigham-Grette and Steve Petsch
Every year, sea ice cover in the Arctic Ocean shrinks to a low point in mid-September. This year it measures just 1.44…

LATEST ARTICLES

green energy2 3
Four Green Hydrogen Opportunities for the Midwest
by Christian Tae
To avert a climate crisis, the Midwest, like the rest of the country, will need to fully decarbonize its economy by…
ug83qrfw
Major Barrier to Demand Response Needs to End
by John Moore, On Earth
If federal regulators do the right thing, electricity customers across the Midwest may soon be able to earn money while…
trees to plant for climate2
Plant These Trees To Improve City Life
by Mike Williams-Rice
A new study establishes live oaks and American sycamores as champions among 17 “super trees” that will help make cities…
north sea sea bed
Why We Must Understand Seabed Geology To Harness The Winds
by Natasha Barlow, Associate Professor of Quaternary Environmental Change, University of Leeds
For any country blessed with easy access to the shallow and windy North Sea, offshore wind will be key to meeting net…
3 wildfire lessons for forest towns as Dixie Fire destroys historic Greenville, California
3 wildfire lessons for forest towns as Dixie Fire destroys historic Greenville, California
by Bart Johnson, Professor of Landscape Architecture, University of Oregon
A wildfire burning in hot, dry mountain forest swept through the Gold Rush town of Greenville, California, on Aug. 4,…
China Can Meet Energy and Climate Goals Capping Coal Power
China Can Meet Energy and Climate Goals Capping Coal Power
by Alvin Lin
At the Leader’s Climate Summit in April, Xi Jinping pledged that China will “strictly control coal-fired power…
Blue water surrounded by dead white grass
Map tracks 30 years of extreme snowmelt across US
by Mikayla Mace-Arizona
A new map of extreme snowmelt events over the last 30 years clarifies the processes that drive rapid melting.
A plane drops red fire retardant on to a forest fire as firefighters parked along a road look up into the orange sky
Model predicts 10-year burst of wildfire, then gradual decline
by Hannah Hickey-U. Washington
A look at the long-term future of wildfires predicts an initial roughly decade-long burst of wildfire activity,…

 Get The Latest By Email

Weekly Magazine Daily Inspiration

New Attitudes - New Possibilities

InnerSelf.comClimateImpactNews.com | InnerPower.net
MightyNatural.com | WholisticPolitics.com | InnerSelf Market
Copyright ©1985 - 2021 InnerSelf Publications. All Rights Reserved.