Science is undergoing a data explosion, and astronomy is leading the way. Modern telescopes produce terabytes of data per observation, and the simulations required to model our observable Universe push supercomputers to their limits. To analyse this data scientists need to be able to think computationally to solve problems. In this course you will investigate the challenges of working with large datasets: how to implement algorithms that work; how to use databases to manage your data; and how to learn from your data with machine learning tools. The focus is on practical skills - all the activities will be done in Python 3, a modern programming language used throughout astronomy. Regardless of whether you’re already a scientist, studying to become one, or just interested in how modern astronomy works ‘under the bonnet’, this course will help you explore astronomy: from planets, to pulsars to black holes. Course outline: Week 1: Thinking about data - Principles of computational thinking - Discovering pulsars in radio images Week 2: Big data makes things slow - How to work out the time complexity of algorithms - Exploring the black holes at the centres of massive galaxies Week 3: Querying data using SQL - How to use databases to analyse your data - Investigating exoplanets in other solar systems Week 4: Managing your data - How to set up databases to manage your data - Exploring the lifecycle of stars in our Galaxy Week 5: Learning from data: regression - Using machine learning tools to investigate your data - Calculating the redshifts of distant galaxies Week 6: Learning from data: classification - Using machine learning tools to classify your data - Investigating different types of galaxies Each week will also have an interview with a data-driven astronomy expert. Note that some knowledge of Python is assumed, including variables, control structures, data structures, functions, and working with files.
Learning from data
Loading...
The University of Sydney의 강좌에서
Data-driven Astronomy
397개의 평가
수업에서
Learning from data: regression
This module introduces the idea of machine learning. We look at standard methodology for running machine learning experiments, and then apply this to calculating redshifts of distant galaxies using decision trees for regression.
강사 만나기
Tara MurphyAssociate Professor
School of Physics
Simon MurphyPostdoctoral Researcher
School of Physics
[MUSIC]
Good classification systems is central to most areas of science and
astronomy is no exception.
As early astronomers started collecting data, they naturally tried to order and
structure it to help make sense of their observations.
One of the most fascinating stories is that of Annie Jump Cannon and
the classification of stellar spectra.
In the late 1800s, Edward Pickering, the director of Harvard College Observatory,
embarked on a survey of the whole sky using the relatively new
technique of stellar spectroscopy.
Over several decades, a team of human computers worked on classifying the stars
into different types, based on the presence of hydrogen,
helium and other absorption lines in their spectra.
During this time, they developed and
experimented with a range of different classification schemes,
before converging on the familiar OBAFGKM scheme we still use today.
Annie Jump Cannon was an extremely productive classifier.
Between 1911 and 1915,
she classified 5,000 stars a month, resulting in an incredible
set of over 200,000 stars that eventually formed the Henry Draper catalogue.
Why is classification so important?
Firstly, classification based purely on observational properties of stars and
galaxies is an empirical result that can survive regardless of changes
in the theoretical interpretations.
Secondly, the amount of information we need to process if we treat each star or
galaxy from a large survey as an individual is so
overwhelming as to be nearly useless.
Classification allows us to draw meaningful conclusions from data
by abstracting it into categories.
Most importantly,
classification can lead to a deeper physical understanding of the data.
In the case of Annie Cannon, the insight she developed from classifying so
many stars led her and the team to develop a new classification system.
Later, another astronomer Cecilia demonstrated that this scheme
reflected a key physical property of stars, their surface temperature.
Despite Annie Cannon's incredible effort,
she wouldn't be able to keep up with today's data rates.
In a world where the space telescope will absorb one billion stars,
a quick back in the envelope tells us that it would take
any of a 16,000 years to classify them by hand.
So what is the answer?
For that, we can skip forward half a century to one of the break through
developments in computer science, the emergence of a failed of machine learning.
Machine learning algorithms are a class of algorithms that can improve
their performance as they're exposed to more data.
In other words they can learn how to do a task better.
There are examples of machine learning everywhere.
Postal mail is automatically sorted using hand writing recognition.
Products or videos are recommended based on your individual tastes and habits.
And computers can now drive cars.
All powered by sophisticated machine learning systems.
In astronomy, machine learning has been used to classify stars and
galaxies, identify volcanos on the surface of Venus, discover pulsars
in massive high-resolution data sets, and many other applications.
From a scientific perspective, some machine learning methods actually mimic
the process followed by human experts quite closely.
This isn't surprising, given the development of artificial neuron network
was inspired by trying to model how the human brain learns.
In this module, we're going to use machine learning to calculate the red shift of
galaxies from their measured colors.
This task is ideal for machine learning and
there has been a substantial amount of research in this area.
Driven by large galaxy surveys such as the Sloan Digital Sky Survey that
observed millions of galaxies.
It's relatively straightforward for a person with a help of a computer to
measure a red shift from a galaxy with an observed spectrum and
hence work out how far away the galaxy is.
But many galaxies have not been observed spectroscopically, we only have images.
In addition, the sheer number of galaxies in these surveys
makes this task impractical to do by hand.
In this module, you'll learn one way of estimating red shifts for
distant galaxies and then train a computer program to calculate red shift for
galaxies it has never seen before.
[MUSIC]
Coursera 카탈로그 살펴보기
무료로 참여해 맞춤화된 추천, 업데이트 및 제안을 받아보세요.
Coursera는 세계 최고의 교육에 대한 보편적인 접근 기회를 제공하며,
최고의 대학 및 기관들과 제휴하여 온라인 강좌를 제공합니다.
© 2019 Coursera Inc. All rights reserved.
