CSCI-B 555 Machine Learning
3 credits
- Prerequisite(s): Programming, calculus, linear algebra, probability, and statistics
- Delivery: On-Campus
- Semesters offered: Fall (Check the schedule to confirm.)
- Equivalent(s): CSCI 57800 Statistical Machine Learning
Description
This course covers the theory and practice of constructing algorithms that learn functions and choose optimal decisions from data and knowledge. Topics include mathematical and probabilistic foundations, MAP classification/regression, linear and logistic regression, neural networks, support vector machines, Bayesian networks, tree models, committee machines, kernel functions, EM, density estimation, accuracy estimation, normalization, and model selection.
Topics
Fundamental algorithms
- Decision trees
- k-nearest neighbors
Model selection
- Model selection
- Experimental design
Linear models
- Linear regression
- Logistic regression
- Optimization
Features
- Feature engineering
- Regularization
Neural networks
- Perceptron
- Backpropagation
- Deep learning
Probabilistic models and reinforcement learning
- Naïve Bayes
- Bayesian networks
- Hidden Markov models
- Reinforcement learning
Dimensionality reduction and clustering
- Dimensionality reduction
- k-means
- Ensemble methods
Applications
- Image recognition
- Machine translation
- Product recommendation
- Sentiment analysis
Implementation and evaluation
- Training models
- Evaluating models
- Applying machine learning to real-world problems
Learning Outcomes
- Analyze and interpret the results from different fundamental algorithms, such as decision trees and k-nearest neighbors. CS 1
- Evaluate the effectiveness of different model selection strategies and experimental designs in various contexts. CS 1
- Design and implement linear models, optimizing them for specific problem sets. CS 4
- Analyze the impact of different feature engineering and regularization techniques on the performance of machine learning models. CS 4
- Develop neural network models and evaluate their performance. CS 4
- Implement probabilistic models and reinforcement learning strategies in solving complex problems. CS 1
- Analyze and design solutions for dimensionality reduction and clustering using various methods, including <em>k</em>-means and ensemble methods. CS 4
- Develop solutions for real-world problems using machine learning applications such as image recognition and sentiment analysis. CS 4
- Develop a comprehensive project demonstrating the ability to implement and evaluate machine learning solutions effectively. CS 7
Policies and Procedures
Please be aware of the following linked policies and procedures. Note that in individual courses instructors will have stipulations specific to their course.