Morphological Classification of Galaxies based on Computer Vision features using CBR and Rule Based Systems

Size: px
Start display at page:

Download "Morphological Classification of Galaxies based on Computer Vision features using CBR and Rule Based Systems"

Transcription

1 Morphological Classification of Galaxies based on Computer Vision features using CBR and Rule Based Systems Devendra Singh Dhami Tasneem Alowaisheq Graduate Candidate School of Informatics and Computing Graduate Candidate School of Informatics and Computing Masters in Computer Science PhD in Computer Science 3801 S. Sharon Dr. 1227, S. Fenbrook Lane Bloomington, Indiana Bloomington, Indiana Abstract There are 10^11 galaxies in the universe and many more are getting discovered at a very fast rate. It becomes really necessary to classify these heavenly bodies into some appropriate classes to help them study further. There have been several approach taken to classify these galaxies with varying degrees of success. In this paper we present two systems: a Case Based Reasoning system and a rule based system designed to perform the classification task. We make use of the morphological features i.e. features describing the structure of the galaxy like its shape, central characteristics (has a bar or bulge at its center) etc., we extract from our Computer Vision System and try to classify the galaxies according to the extracted features. 1. History and Motivation The problem we are tackling is a really interesting one and goes back to the advent of Astronomy. When the first galaxies were discovered, they were very few in number and thus a need of their classification was not a demanding task. But then with the advancement of technology there was a burst in the number of galaxies that were being found and the need of their classification arose. Edwin Hubble classified the galaxies into something called the Hubble sequence in 1936 [1]. This was purely done in a manual way relying on the images that were seen through the telescopes. There were few other systems like the De Vaucoulers syetem and the Yerkes (or Morgan) schemes that were invented to classify the galaxies [1]. Then in 2007, a citizen science project called the Galaxy Zoo was launched, in which millions of volunteers around the world were asked to classify the galaxies based on the images from the Sloan Digital Sky Survey (SDSS) database. The project was successful but with new galaxies being discovered at a really rapid pace, a need was felt to automate the classification process. Some works used Machine learning techniques for the process like the one by Gauci, Adami and Abela [3] that and the one by Banerji [4] were instrumental in introducing the world of Artificial Intelligence to this problem. 2. The Problem We tried to apply the Case Based Reasoning technique to this problem as we figured that it can be very relevant to the problem domain. As the number of galaxies increase we can use the already classified galaxies to classify them. This approach will require the morphological features of the galaxies that can be extracted using Computer Vision techniques. The problem proved to be very interesting and challenging at the same time. Interesting because the Case based reasoning and the rule based method had not been applied before to this domain and to this problem. Challenging because the images of the galaxies were very noisy and small (Figure 1) and the extraction of features was difficult and time-consuming. As the CBR and Rule based system were completely dependent on the extracted features the classification proved to be a hard task. Another problem that was faced was the comparison metric between the old and the new cases. Several approaches were tried some of which performed poorly and some performed satisfactorily. Another challenge was the staggering size of the database. We had around 283 Thousand images and we tested the systems on only 32 images but it can be scaled up to any number of images with none or minor modification. The related work on the area has mainly been on Machine learning classification algorithms and Neural Network systems which proved to be quite effective but were still

2 dependent heavily on the Computer Vision technique. Our system suffers from the same flaw but the results shows that the CBR systems performs relatively well. The CBR approach makes its contribution as this classification process based on morphological features can be made simpler and relatively faster if old examples of the classification can be used as cases for the new galaxies being found. This can reduce the need for a new classification algorithm to be applied every time and although will need a staggering database for the cases will prove to be less costly than other techniques in the long run. This approach will also give us the luxury of having the old data at our disposal and thus the possibility of wrong classification can be significantly reduced. Figure 1. Example of a Galaxy Image. belong to a single class. The class to which they belong will depend on the probabilities of their subclasses. Figure 2. Example of the Human Classified data We build the cases by reading the data from the excel sheet and putting it into the proper case format. We would like to mention here that another challenge we faced during this project was the decision on the case format. A case format for CBR system for human data did not perform well with the CBR system for extracted features and thus we went for two case formats, which although are different but are same in their basic design. We first try to find the class for which the probabilities of the subclasses sum up to 1, as this indicates a 100% probability that the galaxy will belong to that said class. If we don t find such a class then we search for a sum of 0.99 as we feel that this is a big enough probability to safely assume that the galaxy will belong to the class in question. If both the above conditions are not met then we take the nearest neighbor approach. We find out the case for which the distance between the new case and the case is lowest. Then we take the class for which the distance is lowest. This gives us a good enough measure of the class to which the galaxy belongs. Figure 3 shows an example output: 3. Our Approach We implemented 3 systems for this project, 2 Case based reasoning systems and a rule based system. One of the Case Based systems was made for the human data available. By human data we mean the data we obtained from the galaxy Zoo project [2] in which the human volunteers classified the galaxies into various classes. The other Case base and the rule based system were implemented for the features obtained from the computer vision systems. 3.1 CBR system for the human data The format of the human classified data is static. (Figure 2) The data provides us with the galaxy id and the class names with the probabilities as stated by humans of each galaxy belonging to the certain class. We take a very simple approach in this CBR system. We take the liberty of assuming that the galaxies classified by the humans will Figure 3. Output for Classification of Human data

3 3.2 Rule Based System for extracted features The input data we use for the rule based system is the features we output from the Computer vision system. Figure 4 shows an example of the format of the excel sheet we output from the CV system. We also implement a conflict resolution strategy in the rule based system which was required so that a false rule does not fire. Here we check the rule that needs to be fired, which we obtained from the consequent, with the antecedent of the matched rule. We also check whether there is a consequent rule that needs to be fired because not all rules are connected to other rules. We write the classified output in a text file whose snippet is shown below in Figure 7. The first element is the galaxy id and the rest are the classes to which the galaxy belongs to. Figure 4. Example of feature output by CV system for rule based system. This serves as input to the rule based system i.e. goes into the working memory. Figure 6. Running the Rule-Based System The topmost rows is the feature to be detected and the values show the features detected. We have taken care to keep the formats of the rules (Figure 5) in a way that made the task of generating the working memory from the feature output and its comparison with the rules easier. Figure 7. Output of the rule based system 3.3 CBR system for extracted feature Figure 5. Examples of rules in our system The rules consist of the complete rule number as its first part (Rule-1, Rule-2 etc.). This was done because the rule 1 is related to the Rule 2 and Rule 3 to Rule 9 and so on. The second part of the rule gives the antecedent which is the feature of the galaxies in that subclass and the consequent which gives the class number and the Rule which needs to fired next for that galaxy. The CBR system differs with the rule based system in the way the extracted features from the Computer Vision system are sent as input to the system. As shown in Figure 4, for rule based systems we directly send the extracted features to be compared with the rules. In CBR system, we take a little different approach. Here we send the numerical values we received for the parameters of every feature. We do this to test whether the CBR system can find out the features along with the class for each galaxy. Thus we push the CBR system a step further. The CBR system does not disappoint us and is able to search and output the correct feature along with the classes in a greater number of instances. Figure 8 below shows the input file for the CBR system as received from the CV system. The 1 st row indicates the parameters names for different features and their values are indicated in the rows below. ID denotes the galaxy ID.

4 base. Then we normalize all the values for that feature by dividing the feature values with the minimum value. We then find out the nearest case to the normalized value to obtain the feature. A classifier is then used to classify the feature into its class. We write the features obtained for a galaxy in an excel sheet (Figure 11) and the class obtained in a text file (Figure 12). Figure 10. Running the CBR system for features Figure 8. Example of feature output by CV system for CBR system. This serves as input to the CBR system i.e. comprises of the new cases We train our CBR system with selected cases (not among the input cases). Some of the rules are shown in Figure 9 below. Figure 9. Examples of training rules for CBR system We take 2 approaches for the different features, namely case retrieval for the first 3 features is done in one way and that for the last 3 features is done another way. For the first 3 features we use something we call the weight matrix. This is actually a list of lists that consists of weights that are assigned for each features for each case. These weights are assigned according to the concept that the new case should be near to the one of the training cases. The weights are either 1(very near to a case in the knowledge base), 0.5(neither very near nor very far to a case in the knowledge base) or -1(very far from a case in the knowledge base) Then we get the best case according to the weights to get the feature. Then we use a classifier to classify that feature to a class. For the last 3 features we find the minimum value for a feature from the new case and all cases in the knowledge Figure 11. The features as obtained by the CBR system. First row represents the feature name and the values obtained are in the below rows Figure 12. Output of the CBR system In Figure 12, the 1 st column represents the galaxy id and the classes it belongs to are shown after the id.

5 4. Strengths and Weaknesses We feel that our approach has the following strengths: 1. The system is simple in design but efficient. 2. CBR system and the Rule based system give a good estimation of the galaxy classes. 3. System is domain-independent. 4. CBR System is expert knowledge independent. The weaknesses are as follows: 1. Case and rule format is rigid. 2. System not yet tested on a large dataset. The results we obtained so far are not a sufficient measure for accuracy but they provided a good performance indictor 3. Stricter heuristic required for CBR systems. 5. Relation to other works The closest related works we found was by Kasivajhula, Raghavan, and Shah[5]. Other related works were by Gauci, Adami and Abela [3] and Banerji [4] that used Machine learning classification algorithms to classify the galaxies. In [3] decision tree learning algorithms and fuzzy inferencing systems are applied to the problem. The main aim here was to develop distinct models for spiral, elliptical galaxies and stars or unknown space artifacts. For decision tree learning algorithms they use the Classification and Regression Tree (CART) scheme and random forests and for the fuzzy systems they define ifthen rules that deal with fuzzy consequents and antecedants, which is like a rule based system but yet different in a way that they define a degree of membership from 0 to 1 for each class and classify them into bands, namely i and r bands. Their approach is quite different from our approach in this project as they try to find which parameters are important for galaxy classification whereas we use the morphological features for the classification assuming all are important and also look at how similar and distinct each galaxy is to another galaxy. For their Machine learning algorithms they get a 81% accuracy whereas for fuzzy systems they get a 75% accuracy. In [4] an artificial neural network is trained on space objects classified by the human eye (they use the SDSS dataset, which we have also used in this project) and test whether the machine-learning algorithm can reproduce the human results. For classification they use different sets of parameters. One set of input parameters are based on colors and profile fitting. Another set is based on shape and texture as well as the concentration. Their third and final set is a combination of 12 parameters. They classify the galaxies into 3 categories, early types, spirals and artifacts. Using the 1 st set they have a high accuracy of about 87%, 84% and 95% accuracy for the 3 categories. The second set gives around 84%, 87% and 28% accuracy for the 3 categories. The third set gives the highest accuracy, 92%, 92% and 96% for the 3 categories. In [5] they compare 3 classic Machine learning algorithms, Support vector machines, Random Forests and Naïve Bayes in the task of galaxy classification. They follow the same approach as us in extracting the various features of the galaxies. They classify the galaxies into 3 categories, Elliptical, Spiral and Irregular and then subdivide these into 7 classifications, namely, E0, E7, Sa, Sc, SBa, SBc and I. They used 119 images from Zsolt Frei s galaxy catalog[6]. The below Figure 13 (obtained from [5]) shows their classification results. Figure 13. Classification results of [5] Another work was by Godreya and Lolling[7] who described a computational scheme to develop an automatic galaxy classifier. They presented two types of classifiers. One used geometric shape features (which is our approach) and the other used pixel images of galaxies and artificial neural networks. They used the SDSS database. Figure14 below (adopted from 7) shows their computational scheme flowchart.

6 Here is another example for galaxy id The case based system classifies it into ['Class 1.2', 'Class 3.2', 'Class 4.1', 'Class 9.2', 'Class 10.3', 'Class 11.1'] Whereas the rule based system outputs [' Class-4.1 Class-10.3 Class-11.1 Class-1.2 Class-2.2 Class-3.2 Class-9.2'] Again the results are very close to each other but case based system fares better. 7. Future work We feel that our future work consists of expanding our system for a large dataset. Also changes to heuristics for case based system are required for better performance. Another region where we think our system lags is the rigidity of rules and cases. We would want to remove this rigidness. Figure 14. Computational Scheme Flowchart of [7] Their Feature extractor is a bit wider than ours and the classification technique used is Neural Networks. For their Shape feature classifier they used shape properties like elongation, Convexity and area to name among a few. For their direct image based classifier they trained a NN on 215 galaxies and then tested their network on unknown galaxies. For this method their learning probability was 97.16% but the correct identification/classification rate was only 51.35%. 6. Results The results from rule based systems and the case based system go neck to neck in our project. For example for the galaxy id case based system gives the following classes: ['Class 1.1', 'Class 3.2', 'Class 4.2', 'Class 9.2', 'Class 10.2', 'Class 11.6'] The rule based system for the same galaxy gives the following result: [' Class-4.2 Class-11.6 Class-1.1 Class-2.2 Class- 3.2 Class-9.2'] As we can see the results are fairly close to each other with only the Class 10.2 missing from the rule based system and Class 2.2 missing from the case based system. The human data says that the case based system is closer to the human classification. 8. Conclusion We conclude that the case based system fares better than the rule based system in most of the cases we encountered. Though we did not test our systems for a very large number of images but we do get a certain idea of as to how the systems might perform when tested on a large dataset with a few modifications. Case based systems are an interesting application to this domain which we feel is quite relevant for the CBR systems but not much work has been done on this domain. We believe that with a better feature extractor the CBR and rule based systems can prove to be handy tolls for this classification tasks. 9. References [1] ion [2] [3] Gauci, Adam, Kristian Zarb Adami, and John Abela. "Machine Learning for Galaxy Morphology Classification." arxiv preprint arxiv: (2010) [4] Banerji, Manda, Ofer Lahav, Chris J. Lintott, Filipe B. Abdalla, Kevin Schawinski, Steven P. Bamford, Dan Andreescu et al. "Galaxy Zoo: reproducing galaxy morphologies via machine learning." Monthly Notices of the Royal Astronomical Society 406, no. 1 (2010): [5] Kasivajhula, Siddhartha, Naren Raghavan, and Hemal Shah. "Morphological galaxy classification using machine learning." Monthly Notices of the Royal Astronomical Society 8 (2007): 1-8.

7 [6] Frei, Zsolt, Puragra Guhathakurta, James E. Gunn, and J. Anthony Tyson. "A catalog of digital images of 113 nearby galaxies." arxiv preprint astro-ph/ (1995). [7] [7] Goderya, Shaukat N., and Shawn M. Lolling. "Morphological classification of galaxies using computer vision and artificial neural networks: A computational scheme." Astrophysics and Space Science 279, no. 4 (2002): [8] SDSS database

Knowledge-Based Morphological Classification of Galaxies from Vision Features

Knowledge-Based Morphological Classification of Galaxies from Vision Features The AAAI-17 Workshop on Knowledge-Based Techniques for Problem Solving and Reasoning WS-17-12 Knowledge-Based Morphological Classification of Galaxies from Vision Features Devendra Singh Dhami School of

More information

Classifying Galaxy Morphology using Machine Learning

Classifying Galaxy Morphology using Machine Learning Julian Kates-Harbeck, Introduction: Classifying Galaxy Morphology using Machine Learning The goal of this project is to classify galaxy morphologies. Generally, galaxy morphologies fall into one of two

More information

Galaxy Growth and Classification

Galaxy Growth and Classification Observational Astronomy Lab: I-1FS Objectives: First Name: Last Name: Galaxy Growth and Classification To understand the concept of color in astronomy. To be able to classify galaxies based on their morphology

More information

Galaxy Classification and the Hubble Deep Field

Galaxy Classification and the Hubble Deep Field Galaxy Classification and the Hubble Deep Field A. The Hubble Galaxy Classification Scheme Adapted from the UW Astronomy Dept., 1999 Introduction A galaxy is an assembly of between a billion (10 9 ) and

More information

Galaxy Classification

Galaxy Classification Galaxies Galaxies are collections of billons of stars; our home galaxy, the Milky Way, is a typical example. Stars, gas, and interstellar dust orbit the center of the galaxy due to the gravitational attraction

More information

BHS Astronomy: Galaxy Classification and Evolution

BHS Astronomy: Galaxy Classification and Evolution Name Pd Date BHS Astronomy: Galaxy Classification and Evolution This lab comes from http://cosmos.phy.tufts.edu/~zirbel/ast21/homework/hw-8.pdf (Tufts University) The word galaxy, having been used in English

More information

Group Member Names: You may work in groups of two, or you may work alone. Due November 20 in Class!

Group Member Names: You may work in groups of two, or you may work alone. Due November 20 in Class! Galaxy Classification and Their Properties Group Member Names: You may work in groups of two, or you may work alone. Due November 20 in Class! Learning Objectives Classify a collection of galaxies based

More information

AUTOMATIC MORPHOLOGICAL CLASSIFICATION OF GALAXIES. 1. Introduction

AUTOMATIC MORPHOLOGICAL CLASSIFICATION OF GALAXIES. 1. Introduction AUTOMATIC MORPHOLOGICAL CLASSIFICATION OF GALAXIES ZSOLT FREI Institute of Physics, Eötvös University, Budapest, Pázmány P. s. 1/A, H-1117, Hungary; E-mail: frei@alcyone.elte.hu Abstract. Sky-survey projects

More information

Physics Lab #10: Citizen Science - The Galaxy Zoo

Physics Lab #10: Citizen Science - The Galaxy Zoo Physics 10263 Lab #10: Citizen Science - The Galaxy Zoo Introduction Astronomy over the last two decades has been dominated by large sky survey projects. The Sloan Digital Sky Survey was one of the first

More information

Shape Descriptors in Morphological Galaxy Classification

Shape Descriptors in Morphological Galaxy Classification Shape Descriptors in Morphological Galaxy Classification Ishita Dutta 1, S. Banerjee 2 & M. De 3 1&2 Department of Natural Science, West Bengal University of Technology 3 Department of Engineering and

More information

SOURCES AND RESOURCES:

SOURCES AND RESOURCES: A Galactic Zoo Lesson plan for grades K-2 Length of lesson: 1 Class Period (60 minutes) Adapted by: Jesús Aguilar-Landaverde, Environmental Science Institute, February 24, 2012 SOURCES AND RESOURCES: An

More information

GalaxyZoo and the Zooniverse of Astronomy Citizen Science

GalaxyZoo and the Zooniverse of Astronomy Citizen Science April 18, 2014 GalaxyZoo and the Zooniverse of Astronomy Citizen Science Joel R. Primack Distinguished Professor of Physics, University of California, Santa Cruz Director, University of California High-Performance

More information

Galaxy Morphologies with

Galaxy Morphologies with Galaxy Morphologies with Karen Masters ICG, Portsmouth 6.5 years of Galaxy Zoo! July 2007Feb 2009 Feb 2009April 2010 Sept 2009Jan 2010 Karen Masters: Galaxy Zoo, 18th November 2013 Apr 2010Aug 2012 Aug

More information

The Galaxy Zoo Project

The Galaxy Zoo Project Astronomy 201: Cosmology Fall 2009 Prof. Bechtold NAME: The Galaxy Zoo Project 200 points Due: Nov. 23, 2010, in class Professional astronomers often have to search through enormous quantities of data

More information

9.6. Other Components of the Universe. Star Clusters. Types of Galaxies

9.6. Other Components of the Universe. Star Clusters. Types of Galaxies Other Components of the Universe 9.6 The most common type of celestial object astronomers see in space is a star. Most stars appear to be gravitationally bound together into groups, and some groups are

More information

Galaxies and Hubble s Law

Galaxies and Hubble s Law Galaxies and Hubble s Law Some Important History: Charles Messier In the early 19 th century, Charles Messier was hunting for comets, but in the telescopes of the time, identifying comets was difficult

More information

The Milky Way & Galaxies

The Milky Way & Galaxies The Milky Way & Galaxies The Milky Way Appears as a milky band of light across the sky A small telescope reveals that it is composed of many stars (Galileo again!) Our knowledge of the Milky Way comes

More information

INSIDE LAB 9: Classification of Stars and Other Celestial Objects

INSIDE LAB 9: Classification of Stars and Other Celestial Objects INSIDE LAB 9: Classification of Stars and Other Celestial Objects OBJECTIVE: To become familiar with the classification of stars by spectral type, and the classification of celestial objects such as galaxies.

More information

View of the Galaxy from within. Lecture 12: Galaxies. Comparison to an external disk galaxy. Where do we lie in our Galaxy?

View of the Galaxy from within. Lecture 12: Galaxies. Comparison to an external disk galaxy. Where do we lie in our Galaxy? Lecture 12: Galaxies View of the Galaxy from within The Milky Way galaxy Rotation curves and dark matter External galaxies and the Hubble classification scheme Plotting the sky brightness in galactic coordinates,

More information

Galaxies. What is a Galaxy? A bit of History. A bit of History. Three major components: 1. A thin disk consisting of young and intermediate age stars

Galaxies. What is a Galaxy? A bit of History. A bit of History. Three major components: 1. A thin disk consisting of young and intermediate age stars What is a Galaxy? Galaxies A galaxy is a collection of billions of stars, dust, and gas all held together by gravity. Galaxies are scattered throughout the universe. They vary greatly in size and shape.

More information

How Did the Universe Begin?

How Did the Universe Begin? How Did the Universe Begin? As we will discuss in this lecture, it looks like the Universe started about 14 billion years ago and has been expanding (space stretching) ever since. The model of what happened

More information

Galaxy Zoo. Materials Computer Internet connection

Galaxy Zoo. Materials Computer Internet connection Name: Date: Galaxy Zoo Objectives: Distinguish between different types of galaxies Identify the various features of each subclass Contribute data that will be used by astronomers in their work Learn to

More information

Excerpts from previous presentations. Lauren Nicholson CWRU Departments of Astronomy and Physics

Excerpts from previous presentations. Lauren Nicholson CWRU Departments of Astronomy and Physics Excerpts from previous presentations Lauren Nicholson CWRU Departments of Astronomy and Physics Part 1: Review of Sloan Digital Sky Survey and the Galaxy Zoo Project Part 2: Putting it all together Part

More information

Laboratory: Milky Way

Laboratory: Milky Way Department of Physics and Geology Laboratory: Milky Way Astronomy 1402 Equipment Needed Quantity Equipment Needed Quantity Milky Way galaxy Model 1 Ruler 1 1.1 Our Milky Way Part 1: Background Milky Way

More information

SKINAKAS OBSERVATORY. Astronomy Projects for University Students PROJECT GALAXIES

SKINAKAS OBSERVATORY. Astronomy Projects for University Students PROJECT GALAXIES PROJECT 7 GALAXIES Objective: The topics covered in the previous lessons target celestial objects located in our neighbourhood, i.e. objects which are within our own Galaxy. However, the Universe extends

More information

Applied Machine Learning for Design Optimization in Cosmology, Neuroscience, and Drug Discovery

Applied Machine Learning for Design Optimization in Cosmology, Neuroscience, and Drug Discovery Applied Machine Learning for Design Optimization in Cosmology, Neuroscience, and Drug Discovery Barnabas Poczos Machine Learning Department Carnegie Mellon University Machine Learning Technologies and

More information

Some Applications of Machine Learning to Astronomy. Eduardo Bezerra 20/fev/2018

Some Applications of Machine Learning to Astronomy. Eduardo Bezerra 20/fev/2018 Some Applications of Machine Learning to Astronomy Eduardo Bezerra ebezerra@cefet-rj.br 20/fev/2018 Overview 2 Introduction Definition Neural Nets Applications do Astronomy Ads: Machine Learning Course

More information

Figure 19.19: HST photo called Hubble Deep Field.

Figure 19.19: HST photo called Hubble Deep Field. 19.3 Galaxies and the Universe Early civilizations thought that Earth was the center of the universe. In the sixteenth century, we became aware that Earth is a small planet orbiting a medium-sized star.

More information

Data Mining Classification: Basic Concepts and Techniques. Lecture Notes for Chapter 3. Introduction to Data Mining, 2nd Edition

Data Mining Classification: Basic Concepts and Techniques. Lecture Notes for Chapter 3. Introduction to Data Mining, 2nd Edition Data Mining Classification: Basic Concepts and Techniques Lecture Notes for Chapter 3 by Tan, Steinbach, Karpatne, Kumar 1 Classification: Definition Given a collection of records (training set ) Each

More information

Galaxies & Introduction to Cosmology

Galaxies & Introduction to Cosmology Galaxies & Introduction to Cosmology Other Galaxies: How many are there? Hubble Deep Field Project 100 hour exposures over 10 days Covered an area of the sky about 1/100 the size of the full moon Probably

More information

Galaxies. Lecture Topics. Lecture 23. Discovering Galaxies. Galaxy properties. Local Group. History Cepheid variable stars. Classifying galaxies

Galaxies. Lecture Topics. Lecture 23. Discovering Galaxies. Galaxy properties. Local Group. History Cepheid variable stars. Classifying galaxies Galaxies Lecture 23 APOD: NGC 3628 (The Hamburger Galaxy) 1 Lecture Topics Discovering Galaxies History Cepheid variable stars Galaxy properties Classifying galaxies Local Group 2 23-1 Discovering Galaxies

More information

Galaxies and the Universe

Galaxies and the Universe Standard 7.3.1: Recognize and describe that the Sun is a medium-sized star located near the edge of a diskshaped galaxy of stars and that the universe contains many billions of galaxies and each galaxy

More information

THE HUBBLE SEQUENCE. This use case explores the morphology of galaxies and their classification according to the Hubble Sequence.

THE HUBBLE SEQUENCE. This use case explores the morphology of galaxies and their classification according to the Hubble Sequence. THE HUBBLE SEQUENCE G. Iafrate (a), M. Ramella (a) e V. Bologna (b) (a) INAF - Astronomical Observatory of Trieste (b) Istituto Comprensivo S. Giovanni - Sc. Sec. di primo grado M. Codermatz" - Trieste

More information

Figure 1: (Image Credit)

Figure 1: (Image Credit) Ricky Leon Murphy HET609 Semester 1 2005 Project 83 May 30, 2005 Spiral Galaxy Morphology and a Comparison of Four Spiral Galaxies Stars are grouped into island universes called galaxies and there are

More information

Towards a Data-driven Approach to Exploring Galaxy Evolution via Generative Adversarial Networks

Towards a Data-driven Approach to Exploring Galaxy Evolution via Generative Adversarial Networks Towards a Data-driven Approach to Exploring Galaxy Evolution via Generative Adversarial Networks Tian Li tian.li@pku.edu.cn EECS, Peking University Abstract Since laboratory experiments for exploring astrophysical

More information

Automated Classification of Galaxy Zoo Images CS229 Final Report

Automated Classification of Galaxy Zoo Images CS229 Final Report Automated Classification of Galaxy Zoo Images CS229 Final Report 1. Introduction Michael J. Broxton - broxton@stanford.edu The Sloan Digital Sky Survey (SDSS) in an ongoing effort to collect an extensive

More information

THE UNIVERSE CHAPTER 20

THE UNIVERSE CHAPTER 20 THE UNIVERSE CHAPTER 20 THE UNIVERSE UNIVERSE everything physical in and Includes all space, matter, and energy that has existed, now exists, and will exist in the future. How did our universe form, how

More information

LESSON 1. Solar System

LESSON 1. Solar System Astronomy Notes LESSON 1 Solar System 11.1 Structure of the Solar System axis of rotation period of rotation period of revolution ellipse astronomical unit What is the solar system? 11.1 Structure of the

More information

Inferring Galaxy Morphology Through Texture Analysis

Inferring Galaxy Morphology Through Texture Analysis Inferring Galaxy Morphology Through Texture Analysis 1 Galaxy Morphology Kinman Au, Christopher Genovese, Andrew Connolly Galaxies are not static objects; they evolve by interacting with the gas, dust

More information

A Hierarchical Model for Morphological Galaxy Classification

A Hierarchical Model for Morphological Galaxy Classification Proceedings of the Twenty-Sixth International Florida Artificial Intelligence Research Society Conference A Hierarchical Model for Morphological Galaxy Classification Maribel Marin and L. Enrique Sucar

More information

Outline. The Rotation of the Galaxy. Astronomy: The Big Picture

Outline. The Rotation of the Galaxy. Astronomy: The Big Picture I decided to cancel class next week! No homework until you get back. Honor credit need to have those papers soon! Exam 2 Grades are posted. Nighttime observing should be posted today. Interest in grade

More information

Galaxies. Introduction. Different Types of Galaxy. Teacher s Notes. Shape. 1. Download these notes at

Galaxies. Introduction. Different Types of Galaxy. Teacher s Notes. Shape. 1. Download these notes at 1. Introduction A galaxy is a collection of stars, the remains of stars, gas and dust, and the mysterious dark matter. There are many different types and sizes of galaxies, ranging from dwarf galaxies

More information

The Milky Way Galaxy (ch. 23)

The Milky Way Galaxy (ch. 23) The Milky Way Galaxy (ch. 23) [Exceptions: We won t discuss sec. 23.7 (Galactic Center) much in class, but read it there will probably be a question or a few on it. In following lecture outline, numbers

More information

April 11, Astronomy Notes Chapter 16.notebook. Types of Galaxies

April 11, Astronomy Notes Chapter 16.notebook. Types of Galaxies The Milky Way is just one of about 50 billion galaxies that are thought to exist. Just as stars can be classified using an H R diagram, galaxies can also be classified according to certain physical properties.

More information

What is the solar system?

What is the solar system? Notes Astronomy What is the solar system? 11.1 Structure of the Solar System Our solar system includes planets and dwarf planets, their moons, a star called the Sun, asteroids and comets. Planets, dwarf

More information

CS 188: Artificial Intelligence Spring Announcements

CS 188: Artificial Intelligence Spring Announcements CS 188: Artificial Intelligence Spring 2010 Lecture 22: Nearest Neighbors, Kernels 4/18/2011 Pieter Abbeel UC Berkeley Slides adapted from Dan Klein Announcements On-going: contest (optional and FUN!)

More information

LET S MAKE A PORTRAIT OF A GALAXY

LET S MAKE A PORTRAIT OF A GALAXY LET S MAKE A PORTRAIT OF A GALAXY Veselka S. Radeva EAAE Summer School Working Group (Bulgaria) Abstract The goal of the exercise is for students to learn about the different types of galaxies, about the

More information

The Neighbors Looking outward from the Sun s location in the Milky Way, we can see a variety of other galaxies:

The Neighbors Looking outward from the Sun s location in the Milky Way, we can see a variety of other galaxies: Galaxies The Neighbors Looking outward from the Sun s location in the Milky Way, we can see a variety of other galaxies: Small Magellanic Cloud (Digital Sky Survey) Large Magellanic Cloud (credit: Eckhard

More information

Physics Lab #9: Measuring the Hubble Constant

Physics Lab #9: Measuring the Hubble Constant Physics 10263 Lab #9: Measuring the Hubble Constant Introduction In the 1920 s, Edwin Hubble discovered a relationship that is now known as Hubble s Law. It states that the recession velocity of a galaxy

More information

Bayesian Classifiers and Probability Estimation. Vassilis Athitsos CSE 4308/5360: Artificial Intelligence I University of Texas at Arlington

Bayesian Classifiers and Probability Estimation. Vassilis Athitsos CSE 4308/5360: Artificial Intelligence I University of Texas at Arlington Bayesian Classifiers and Probability Estimation Vassilis Athitsos CSE 4308/5360: Artificial Intelligence I University of Texas at Arlington 1 Data Space Suppose that we have a classification problem The

More information

Review of Lecture 15 3/17/10. Lecture 15: Dark Matter and the Cosmic Web (plus Gamma Ray Bursts) Prof. Tom Megeath

Review of Lecture 15 3/17/10. Lecture 15: Dark Matter and the Cosmic Web (plus Gamma Ray Bursts) Prof. Tom Megeath Lecture 15: Dark Matter and the Cosmic Web (plus Gamma Ray Bursts) Prof. Tom Megeath A2020 Disk Component: stars of all ages, many gas clouds Review of Lecture 15 Spheroidal Component: bulge & halo, old

More information

THE HUBBLE SEQUENCE. Information and contacts: -

THE HUBBLE SEQUENCE. Information and contacts:  - THE HUBBLE SEQUENCE G. Iafrate (a), M. Ramella (a) e V. Bologna (b) (a) INAF - Astronomical Observatory of Trieste (b) Istituto Comprensivo S. Giovanni - Sc. Sec. di primo grado M. Codermatz" - Trieste

More information

COMA CLUSTER OF GALAXIES

COMA CLUSTER OF GALAXIES COMA CLUSTER OF GALAXIES Learn the basics of galaxy classification and grouping, using actual Hubble Space Telescope images. Keely Finkelstein, McDonald Observatory Curriculum topic Galaxies Big idea of

More information

Holdout and Cross-Validation Methods Overfitting Avoidance

Holdout and Cross-Validation Methods Overfitting Avoidance Holdout and Cross-Validation Methods Overfitting Avoidance Decision Trees Reduce error pruning Cost-complexity pruning Neural Networks Early stopping Adjusting Regularizers via Cross-Validation Nearest

More information

A Spatial-Color Layout Feature for Content-based Galaxy Image Retrieval

A Spatial-Color Layout Feature for Content-based Galaxy Image Retrieval A Spatial-Color Layout Feature for Content-based Galaxy Image Retrieval Yin Cui, Yongzhou Xiang, Kun Rong, Rogerio Feris, Liangliang Cao Department of Electrical Engineering, Columbia University IBM T.

More information

2. Can observe radio waves from the nucleus see a strong radio source there Sagittarius A* or Sgr A*.

2. Can observe radio waves from the nucleus see a strong radio source there Sagittarius A* or Sgr A*. 7/7 The Nucleus of the MW its center 1. Can t see the nucleus in visible light too much stuff in the way. 2. Can observe radio waves from the nucleus see a strong radio source there Sagittarius A* or Sgr

More information

Galaxies The Hubble Sequence Different Types of Galaxies 4 broad Morphological Types created by Edwin Hubble Galaxies come is a variety of shapes and

Galaxies The Hubble Sequence Different Types of Galaxies 4 broad Morphological Types created by Edwin Hubble Galaxies come is a variety of shapes and Galaxies The Hubble Sequence Different Types of Galaxies 4 broad Morphological Types created by Edwin Hubble Galaxies come is a variety of shapes and sizes Edwin Hubble classified the galaxies into four

More information

Tuesday, Thursday 2:30-3:45 pm. Astronomy 100. Tom Burbine

Tuesday, Thursday 2:30-3:45 pm.   Astronomy 100. Tom Burbine Astronomy 100 Tuesday, Thursday 2:30-3:45 pm Tom Burbine tburbine@mtholyoke.edu www.xanga.com/astronomy100 OWL assignment (Due Today) There is be an OWL assignment due on Thursday April 14 at 11:59 pm.

More information

Chapter 30. Galaxies and the Universe. Chapter 30:

Chapter 30. Galaxies and the Universe. Chapter 30: Chapter 30 Galaxies and the Universe Chapter 30: Galaxies and the Universe Chapter 30.1: Stars with varying light output allowed astronomers to map the Milky Way, which has a halo, spiral arm, and a massive

More information

PH104 Lab 2 Measuring Distances Pre-Lab

PH104 Lab 2 Measuring Distances Pre-Lab Name: Lab Time: PH04 Lab 2 Measuring Distances Pre-Lab 2. Goals This is the second lab. Like the first lab this lab does not seem to be part of a complete sequence of the study of astronomy, but it will

More information

Modern Image Processing Techniques in Astronomical Sky Surveys

Modern Image Processing Techniques in Astronomical Sky Surveys Modern Image Processing Techniques in Astronomical Sky Surveys Items of the PhD thesis József Varga Astronomy MSc Eötvös Loránd University, Faculty of Science PhD School of Physics, Programme of Particle

More information

9. High-level processing (astronomical data analysis)

9. High-level processing (astronomical data analysis) Master ISTI / PARI / IV Introduction to Astronomical Image Processing 9. High-level processing (astronomical data analysis) André Jalobeanu LSIIT / MIV / PASEO group Jan. 2006 lsiit-miv.u-strasbg.fr/paseo

More information

A Spatial-Color Layout Feature for Representing Galaxy Images

A Spatial-Color Layout Feature for Representing Galaxy Images A Spatial-Color Layout Feature for Representing Galaxy Images Yin Cui, Yongzhou Xiang, Kun Rong, Rogerio Feris, Liangliang Cao Department of Electrical Engineering, Columbia University IBM T. J. Watson

More information

18.9 SUPPORT VECTOR MACHINES

18.9 SUPPORT VECTOR MACHINES 744 Chapter 8. Learning from Examples is the fact that each regression problem will be easier to solve, because it involves only the examples with nonzero weight the examples whose kernels overlap the

More information

Galaxies. Need a (physically) meaningful way of describing the relevant properties of a galaxy.

Galaxies. Need a (physically) meaningful way of describing the relevant properties of a galaxy. Galaxies Aim to understand the characteristics of galaxies, how they have evolved in time, and how they depend on environment (location in space), size, mass, etc. Need a (physically) meaningful way of

More information

Practice Test: ES-5 Galaxies

Practice Test: ES-5 Galaxies Class: Date: Practice Test: ES-5 Galaxies Multiple Choice Identify the letter of the choice that best completes the statement or answers the question. Light Years * The distance between stars and galaxies

More information

GALAXIES. Hello Mission Team members. Today our mission is to learn about galaxies.

GALAXIES. Hello Mission Team members. Today our mission is to learn about galaxies. GALAXIES Discussion Hello Mission Team members. Today our mission is to learn about galaxies. (Intro slide- 1) Galaxies span a vast range of properties, from dwarf galaxies with a few million stars barely

More information

Major Review: A very dense article" Dawes Review 4: Spiral Structures in Disc Galaxies; C. Dobbs and J Baba arxiv "

Major Review: A very dense article Dawes Review 4: Spiral Structures in Disc Galaxies; C. Dobbs and J Baba arxiv The Components of a Spiral Galaxy-a Bit of a Review- See MBW chap 11! we have discussed this in the context of the Milky Way" Disks:" Rotationally supported, lots of gas, dust, star formation occurs in

More information

Lecture #21: Plan. Normal Galaxies. Classification Properties Distances

Lecture #21: Plan. Normal Galaxies. Classification Properties Distances Lecture #21: Plan Normal Galaxies Classification Properties Distances Messier 31 = M31 Early 20 th Century The Great Debate (4/26/1920): Harlow Shapley (Mt Wilson) vs Heber Curtis (Lick Observatory) Smithsonian

More information

Module 3: Astronomy The Universe Topic 2 Content: The Milky Way Galaxy Presentation Notes

Module 3: Astronomy The Universe Topic 2 Content: The Milky Way Galaxy Presentation Notes On a clear night, you can go outside and view the Moon and the stars scattered throughout the night sky. At times, you can also see neighboring planets. When you look at the sky and these objects, almost

More information

Surprise Detection in Science Data Streams Kirk Borne Dept of Computational & Data Sciences George Mason University

Surprise Detection in Science Data Streams Kirk Borne Dept of Computational & Data Sciences George Mason University Surprise Detection in Science Data Streams Kirk Borne Dept of Computational & Data Sciences George Mason University kborne@gmu.edu, http://classweb.gmu.edu/kborne/ Outline Astroinformatics Example Application:

More information

Lab 1. EXCEL plus some basic concepts such as scientific notation, order of magnitude, logarithms, and unit conversions

Lab 1. EXCEL plus some basic concepts such as scientific notation, order of magnitude, logarithms, and unit conversions COMPUTER LAB 1 EARTH SYSTEMS SCIENCE I PG250 Fall 2010 Hunter College Lab 1. EXCEL plus some basic concepts such as scientific notation, order of magnitude, logarithms, and unit conversions Low Impact

More information

How did the universe form? 1 and 2

How did the universe form? 1 and 2 Galaxies How did the universe form? 1 and 2 Galaxies Astronomers estimate that 40 billion galaxies exist in the observable universe The universe may contain over 100 billion galaxies Even a modest-sized

More information

Machine Learning Methods for Radio Host Cross-Identification with Crowdsourced Labels

Machine Learning Methods for Radio Host Cross-Identification with Crowdsourced Labels Machine Learning Methods for Radio Host Cross-Identification with Crowdsourced Labels Matthew Alger (ANU), Julie Banfield (ANU/WSU), Cheng Soon Ong (Data61/ANU), Ivy Wong (ICRAR/UWA) Slides: http://www.mso.anu.edu.au/~alger/sparcs-vii

More information

Outline Challenges of Massive Data Combining approaches Application: Event Detection for Astronomical Data Conclusion. Abstract

Outline Challenges of Massive Data Combining approaches Application: Event Detection for Astronomical Data Conclusion. Abstract Abstract The analysis of extremely large, complex datasets is becoming an increasingly important task in the analysis of scientific data. This trend is especially prevalent in astronomy, as large-scale

More information

Machine Learning Applications in Astronomy

Machine Learning Applications in Astronomy Machine Learning Applications in Astronomy Umaa Rebbapragada, Ph.D. Machine Learning and Instrument Autonomy Group Big Data Task Force November 1, 2017 Research described in this presentation was carried

More information

From the Big Bang to Big Data. Ofer Lahav (UCL)

From the Big Bang to Big Data. Ofer Lahav (UCL) From the Big Bang to Big Data Ofer Lahav (UCL) 1 Outline What is Big Data? What does it mean to computer scientists vs physicists? The Alan Turing Institute Machine learning examples from Astronomy The

More information

ASTRO504 Extragalactic Astronomy. 2. Classification

ASTRO504 Extragalactic Astronomy. 2. Classification ASTRO504 Extragalactic Astronomy 2. Classification Morphological classification Elliptical (E) galaxies Lenticular (SO) galaxies Spiral (S) galaxies Irregular (Im) galaxies The realm of nebulae Hubble

More information

Homework on Properties of Galaxies in the Hubble Deep Field Name: Due: Friday, April 8 30 points Prof. Rieke & TA Melissa Halford

Homework on Properties of Galaxies in the Hubble Deep Field Name: Due: Friday, April 8 30 points Prof. Rieke & TA Melissa Halford Homework on Properties of Galaxies in the Hubble Deep Field Name: Due: Friday, April 8 30 points Prof. Rieke & TA Melissa Halford You are going to work with some famous astronomical data in this homework.

More information

Fast Hierarchical Clustering from the Baire Distance

Fast Hierarchical Clustering from the Baire Distance Fast Hierarchical Clustering from the Baire Distance Pedro Contreras 1 and Fionn Murtagh 1,2 1 Department of Computer Science. Royal Holloway, University of London. 57 Egham Hill. Egham TW20 OEX, England.

More information

Combining Human and Machine Learning for Morphological Analysis of Galaxy Images

Combining Human and Machine Learning for Morphological Analysis of Galaxy Images PUBLICATIONS OF THE ASTRONOMICAL SOCIETY OF THE PACIFIC, 126:959 967, 2014 October 2014. The Astronomical Society of the Pacific. All rights reserved. Printed in U.S.A. Combining Human and Machine Learning

More information

Data Release 5. Sky coverage of imaging data in the DR5

Data Release 5. Sky coverage of imaging data in the DR5 Data Release 5 The Sloan Digital Sky Survey has released its fifth Data Release (DR5). The spatial coverage of DR5 is about 20% larger than that of DR4. The photometric data in DR5 are based on five band

More information

Homework #7: Properties of Galaxies in the Hubble Deep Field Name: Due: Friday, October points Profs. Rieke

Homework #7: Properties of Galaxies in the Hubble Deep Field Name: Due: Friday, October points Profs. Rieke Homework #7: Properties of Galaxies in the Hubble Deep Field Name: Due: Friday, October 31 30 points Profs. Rieke You are going to work with some famous astronomical data in this homework. The image data

More information

Kyle Lackey PHYS

Kyle Lackey PHYS Kyle Lackey PHYS 730 9-23-15 Galaxies Large systems of gas, dust, stars, and dark matter orbiting around a common center of mass. We estimate that roughly 185 billion galaxies exist within the observable

More information

The Secrets of Galaxies. Student s Guide Advanced Level CESAR s Science Case

The Secrets of Galaxies. Student s Guide Advanced Level CESAR s Science Case The Secrets of Galaxies Student s Guide Advanced Level Introduction Galaxies are fundamental building blocks of the Universe. Some are simple, while others are very complex in structure; some have enormous

More information

It is about 100,000 ly across, 2,000 ly thick, and our solar system is located 26,000 ly away from the center of the galaxy.

It is about 100,000 ly across, 2,000 ly thick, and our solar system is located 26,000 ly away from the center of the galaxy. The Galaxies The Milky Way Galaxy Is a spiral galaxy in which our solar system is located. The center of the galaxy lies in the Sagittarius Constellation. It is about 100,000 ly across, 2,000 ly thick,

More information

Earth Science Lesson Plan Quarter 4, Week 9, Day 1

Earth Science Lesson Plan Quarter 4, Week 9, Day 1 Earth Science Lesson Plan Quarter 4, Week 9, Day 1 Outcomes for Today Standard Focus: Earth Sciences 2.a Students know that the solar system is located on the outer edge of the disk-shaped Milky Way galaxy,

More information

CSE 417T: Introduction to Machine Learning. Final Review. Henry Chai 12/4/18

CSE 417T: Introduction to Machine Learning. Final Review. Henry Chai 12/4/18 CSE 417T: Introduction to Machine Learning Final Review Henry Chai 12/4/18 Overfitting Overfitting is fitting the training data more than is warranted Fitting noise rather than signal 2 Estimating! "#$

More information

Geometric View of Machine Learning Nearest Neighbor Classification. Slides adapted from Prof. Carpuat

Geometric View of Machine Learning Nearest Neighbor Classification. Slides adapted from Prof. Carpuat Geometric View of Machine Learning Nearest Neighbor Classification Slides adapted from Prof. Carpuat What we know so far Decision Trees What is a decision tree, and how to induce it from data Fundamental

More information

Astronomers push the edges of final frontier, find farthest galaxy

Astronomers push the edges of final frontier, find farthest galaxy Astronomers push the edges of final frontier, find farthest galaxy By Associated Press, adapted by Newsela staff on 03.14.16 Word Count 688 This image provided by the Space Telescope Science Institute,

More information

Astronomers discover an active, bright galaxy "in its infancy"

Astronomers discover an active, bright galaxy in its infancy Astronomers discover an active, bright galaxy "in its infancy" By Los Angeles Times, adapted by Newsela staff on 05.18.15 Word Count 825 The galaxy EGS-zs8-1, the most distant galaxy yet seen, was discovered

More information

Galaxy Zoo: the independence of morphology and colour

Galaxy Zoo: the independence of morphology and colour Galaxy Zoo: the independence of morphology and colour Steven Bamford University of Portsmouth / University of Nottingham Chris Lintott, Kevin Schawinski, Kate Land, Anze Slosar, Daniel Thomas, Bob Nichol,

More information

Click Prediction and Preference Ranking of RSS Feeds

Click Prediction and Preference Ranking of RSS Feeds Click Prediction and Preference Ranking of RSS Feeds 1 Introduction December 11, 2009 Steven Wu RSS (Really Simple Syndication) is a family of data formats used to publish frequently updated works. RSS

More information

An analogy. "Galaxies" can be compared to "cities" What would you like to know about cities? What would you need to be able to answer these questions?

An analogy. Galaxies can be compared to cities What would you like to know about cities? What would you need to be able to answer these questions? An analogy "Galaxies" can be compared to "cities" What would you like to know about cities? how does your own city look like? how big is it? what is its population? history? how did it develop? how does

More information

Galaxies. Say Thanks to the Authors Click (No sign in required)

Galaxies. Say Thanks to the Authors Click  (No sign in required) Galaxies Say Thanks to the Authors Click http://www.ck12.org/saythanks (No sign in required) To access a customizable version of this book, as well as other interactive content, visit www.ck12.org CK-12

More information

Parts 3-6 are EXAMPLES for cse634

Parts 3-6 are EXAMPLES for cse634 1 Parts 3-6 are EXAMPLES for cse634 FINAL TEST CSE 352 ARTIFICIAL INTELLIGENCE Fall 2008 There are 6 pages in this exam. Please make sure you have all of them INTRODUCTION Philosophical AI Questions Q1.

More information

UNIVERSITY of PENNSYLVANIA CIS 520: Machine Learning Final, Fall 2013

UNIVERSITY of PENNSYLVANIA CIS 520: Machine Learning Final, Fall 2013 UNIVERSITY of PENNSYLVANIA CIS 520: Machine Learning Final, Fall 2013 Exam policy: This exam allows two one-page, two-sided cheat sheets; No other materials. Time: 2 hours. Be sure to write your name and

More information

Learning algorithms at the service of WISE survey

Learning algorithms at the service of WISE survey Katarzyna Ma lek 1,2,3, T. Krakowski 1, M. Bilicki 4,3, A. Pollo 1,5,3, A. Solarz 2,3, M. Krupa 5,3, A. Kurcz 5,3, W. Hellwing 6,3, J. Peacock 7, T. Jarrett 4 1 National Centre for Nuclear Research, ul.

More information

Introduction to Machine Learning Midterm Exam

Introduction to Machine Learning Midterm Exam 10-701 Introduction to Machine Learning Midterm Exam Instructors: Eric Xing, Ziv Bar-Joseph 17 November, 2015 There are 11 questions, for a total of 100 points. This exam is open book, open notes, but

More information

Virtual Beach Building a GBM Model

Virtual Beach Building a GBM Model Virtual Beach 3.0.6 Building a GBM Model Building, Evaluating and Validating Anytime Nowcast Models In this module you will learn how to: A. Build and evaluate an anytime GBM model B. Optimize a GBM model

More information