Machine learning methods are commonly used across engineering and sciences, from computer systems to physics. Moreover, commercial sites such as search engines, recommender systems (e.g., Netflix, Amazon), advertisers, and financial institutions employ machine learning algorithms for content recommendation, predicting customer behavior, compliance, or risk. As a discipline, machine learning tries to design and understand computer programs that learn from experience for the purpose of prediction or control. In this course, you will learn about principles and algorithms for turning training data into effective automated predictions. We will cover concepts such as representation, over-fitting, regularization, and generalization; topics such as clustering, classification, recommender problems, probabilistic modeling, reinforcement learning; and methods such as on-line algorithms, support vector machines, neural networks/deep learning, hidden Markov models, and Bayesian networks.
(Students in 6.862, the graduate version, should also check http://courses.csail.mit.edu/6.862/)
Midterm on Thursday March 23, closed book, 1.5h, in-class, 20% of grade. If you have a conflict with the exam time, please notify us of your midterm conflict at least two weeks before the exam. The conflict exam (with permission only) will be the next day (Friday, March 24). In exceptional cases, the weight of the midterm can be shifted to the final exam.
Final exam on Thursday, May 25 from 9:00 to 12:00 noon in Track, exam is closed book, 3h, 40% of grade. Any conflicts with the final exam time are handled by the registrar's office.
5 regular homework assignments (10% of the grade in total).
Three projects, each worth 10% of the grade, involving programming in python.
Late submission policy for homeworks and projects:
Assignments are due 9am on the specified Fridays and solutions will be posted immediately after the deadline. Late assignments will NOT be accepted. If you have a valid reason (= dean's note), you must notify us via email. For students who do obtain a dean's note we will add the weight of the missed assignment to the final exam weight. Missed or late assignment without a valid reason correspond to getting the score zero on the assignment.
Collaboration policy for homeworks and projects:
We encourage students to discuss assignments except homework 0 with other students and with the teaching staff to better understand the concepts. When you submit an assignment under your name, we assume that you are certifying that the details are entirely your own work and that you played at least a substantial role in the conception stage. Homework 0 (background) must be completed by each student individually without collaboration.
You should not use results (solutions, code) from other students from this year or previous years in preparing your solutions to any assignment unless you developed the materials while working with other studens in class. Students should never share solutions (or staff solutions) with other students.