STAT598 – Statistical Machine Learning, Exercises of Artificial Intelligence

Download STAT598 – Statistical Machine Learning and more Exercises Artificial Intelligence in PDF only on Docsity! STAT598 – Statistical Machine Learning Course Description This course provides a comprehensive treatment of modern statistical machine learning topics. These include linear regression, kernel method, trees, boosting, and deep neural networks. We understand these subjects from a statistical perspective with some mathematical rigorousness. This course is also a 14-week fully online course. Each week includes 3 lectures. Prerequisite Knowledge of master level of probability and mathematical statistics is required. Basic programming skills using Python is required. Course Outcomes • CO1: Understand classical supervised learning methods such as linear regression, nonlinear regression, random forest, and boosting • CO2: Understand classical unsupervised learning methods such as K-means clustering and PCA • CO3: Understand modern AI techniques using deep neural networks, such as manifold learning and deep generative models • CO4: Practice complex and high dimensional data analysis using the methodology and software packages Technical Requirements The following information has been provided to assist you in preparing to use technology successfully. • A reliable internet connection - capable of consistently streaming video and stable enough to finish short exams without dropping connection. • Access to Purdue's Brightspace Learning Management System - All course content, course readings, and exams will be accessed online through Brightspace. Learning Resources & Texts Required Textbook: 1. The Elements of Statistical Learning by Hastie, Tibshirani, Friedman (Its pdf is available online) Suggested Textbook: 1 Deep Learning by Goodfellow, Bengio, Courville 2 Deep Learning with PyTorch by Stevens, Antiga, Viehmann Instructor’s Online Hours I will be available and respond to student questions as soon as I am available (generally 48) hours during the M-F work week. Student inquiries made during the weekend may experience a delayed response time. Questions about the course content, assignments, or lectures should be asked in the Course Q & A forum provided on the discussion boards. Students are encouraged to answer the questions their peers ask on the Course Q & A forum. Email should only be used for personal questions. When emailing me, please place the course number in the subject line of the email. This will help me tremendously in locating your emails quicker.Virtual Office Hours Virtual Office Hours are a synchronous session through Webex to discuss questions related to the course content. Please check the course site on Brightspace for detailed information about the virtual office hours. Assignments Homework: There will be three written homework assignments to be collected and graded. Late homework and email homework will NOT be accepted without prior permission. Details on these assignments, and guidelines on discussion participation and evaluation will be posted on Brightspace. The due dates for the assignments posted on Brightspace are in Eastern Standard Time (the local time zone of West Lafayette, Indiana). Final Project: The final project is a written statistical report on some interesting data obtained by the student, which will give students an opportunity to review the statistical methodologies covered in the course and the steps taken in the research process. The final project must be typed, double spaced and should be 8-10 pages in length (including tables and graphics). Late projects will be graded down one letter grade. Assignments Points Assignment 1 Assignment 2 Assignment 3 Assignment 3 Final Project 15 % 15 % 15 % 15 % 40 % Total 100% Participation and Assignment policies Some collaboration on homework is acceptable, but each student must do his/her own write-up of the solution to show their full understanding. Direct copying of another student's solution will result in grade of zero for both students. Please remember that an illustration of effort is in itself a creditable achievement. We will be awarding partial credits for all problems depending on approach and degree of completeness. Residential students are expected to attend the lectures in-person, complete homework and final project on time. Online students are expected to watch the lecture videos, complete homework and final project on time. Late submission will cause late penalty on grades. Please check the requirements for each of the assignment on the course site in Brightspace.