Definitions
Linear Neural Networks
- Perceptron
- Adaline
- Stochastic Gradient Descent
Classifiers
- Logistic Regression
- Support Vector Machines
- Decision Trees
- K-Neighbors
Supervised Learning
Where a program is “trained” on a pre-defined dataset. Based off its training data the program can make accurate decisions when given new data.
Unsupervised Learning
Where a program, given a dataset, can automatically find patterns and relationships in that dataset.
Classification
Classification is the supervised learning process of taking some sort of input and assigning a label to it. Generally a binary "yes or no".
Neural Networks
Interconnected nodes that make up a model, used to estimate or approximate functions that depend on a large number of inputs.
Neural networks are usually used when the volume of inputs is far too large for standard machine learning approaches.
Accuracy
The number of correct classification predictions divided by the total number of predictions
Activation Function
A function that enables neural networks to learn nonlinear (complex) relationships between features and the label i.e. ReLu, Sigmoid, etc
Feature (or attribute)
An input variable to a machine learning model. Characteristics pertaining to individuals (i.e. petal length of iris flowers)
Confusion Matrix
An NxN table that summarizes the number of correct and incorrect predictions that a classification model made.
Convolutional Layer
Operations on an input matrix. i.e. convolutional operations on 3x3 slices of a 5x5 matrix
Decision Boundary
The separator between classes learned by a model in a binary class or multi-class classification problems
Epoch
A full training pass over the entire training set such that each example has been processed once.
Gradient Descent
A mathematical technique to minimize loss. Gradient descent iteratively adjusts weights and biases, gradually finding the best combination to minimize loss.
Rectified Linear Unit (ReLU)
An activation function with the following behavior:
If input is negative or zero, then the output is 0.
If input is positive, then the output is equal to the input.
Sigmoid Function
A mathematical function that "squishes" an input value into a constrained range, typically 0 to 1 or -1 to +1
Validation Set
The subset of the dataset that performs initial evaluation against a trained model
Typically, you evaluate the trained model against the validation set several times before evaluating the model against the test set.
Weight
A value that a model multiplies by another value. Training is the process of determining a model's ideal weights; inference is the process of using those learned weights to make predictions.
Course Description
This course is intended for all graduate students and provides an introduction to galaxies and extragalactic astronomy. The emphasis is on making the connection between what we can observe about galaxies and the more fundamental physical properties that we would like to know about the galaxies themselves. We will cover the main techniques used for observing galaxies (imaging, spectroscopy, distance methods, extragalactic surveys), the observed properties (morphology, colors, luminosity functions, surface brightness profiles, kinematics, scaling relations, and spectral energy distributions), and the basic components that make up individual galaxies (stars, gas, dust, central black holes, dark matter) like our own Milky Way. The companion class, ASTR 616 - Galaxies II, builds on this foundation to explore the underlying principles of galaxy formation, galaxy evolution, and cosmology.