Instructor:  Milos Hauskrecht
Computer Science Department
5329 Sennott Square
phone: x4-8845
e-mail: milos_at_cs_pitt_edu
office hours: by appointment

• The final project for the class: problem statement. Timeline:
  • proposals due on April 7, 2020
  • presentations on April 23, 2020 at 9:15am.
  • reports due on April 24, 2020
• The final project presentations will be held on April 23, 2020 at 9:15am on Zoom. The presentations should be 12 minutes + 3 minutes for questions. Please be ready to present by sharing your presentation screen in Zoom. Also, please submit your ppt presentations by 9:00am on April 23, 2020 via CourseWeb. The presentation order is:
  • 9:10am Zoom is activated
  • 9:15am Li Xiaoting (lead)
  • 9:30am Yan Muheng (lead)
  • 9:45am Joe Zapitelli (lead)
  • 10:00am Tristan Maidment (lead)
  • 10:15am Jun Luo (lead)
• A brief description of MIMIC-III
• The course will resume the week of March 23, 2020 and will be held online. The plan is:
  • The course for the remainder of the semester will cover reinforcement learning.
  • Online content and lectures that are available online will be used to cover the topics. The links will be available one day prior to the regularly scheduled class time. No student presentations are expected.
  • Question/answering sessions related to the covered topics will be held online in real time on Zoom platform starting at the regular class time: 9:15am on Tuesdays/Thursdays (Q/A sessions will start on March 26, 2020). The zoom link to the session is now available on the CourseWeb.
  • Short abstract/summaries of readings/videos for each class should be now submitted via CourseWeb.
  • Office hours will be held online on Zoom platform during the regularly scheduled office hour times.
• Groups for the midterm project:
  • Group 1:
    • Joe Zapittelli
    • Xiaoting Li
    • Mesut Unal
  • Group 2:
    • Jun Luo
    • Sarkar Sanchayan
    • Ahmad Diab
  • Group 3:
    • Anthony Sicilia
    • Tristan Maidment
    • Muheng Yan
• Assignment: Please complete the following steps to get access to data for the midterm project.
• Lecture readings At the beginning of every lecture please submit a short summary of the paper/chapter on the reading list for that lecture. The summary should be 2-3 paragraphs long.
• Syllabus
• Please read: Guidelines for preparing your class presentations. .
• Upcoming presentation schedule:
  Low-dimensional data representations, probabilistic models of data:
  • January 28, 2020 Mesut Unal: PCA, autoencoders ( 2018 slides link)
  • January 30, 2020 Xiaoting Li: Latent variable generative models I: pPCA, factor analysis ( 2018 slides link)
  • February 4, 2020 Ahmad Diab: Latent variable generative models II: CVQ, NOCA ( 2018 slides link)
  • February 6, 2020 Jun Luo: Modern generative models of data: Restricted Boltzman machines, Variational autoencoders ( 2018 slides link)
  • February 11, 2020 Tristan Maidment: Generative adversarial networks ( 2018 slides link)
  Document models
  • February 13, 2020 Anthony Sicilia: Singular value decomposition. Applications to latent semantic indexing (LSI), network analysis ( 2018 slides link)
  • February 18, 2020 Joe Zapatelli: Probabilistic LSI, Latent Dirichlet allocation ( 2018 slides link)
  Word models, word similarity
  • February 20, 2020 Muheng Yan : Word projections: CBOW, skipgram, graph-based models ( 2018 slides link)
  Time series/Sequence models:
  • February 25, 2020 Yanbing Xue : Discrete state models: Markov models, HMMs (Yanbing Xue)
  • February 27, 2020 Sarkar Sanchayan : Continuous state models: Autoregressive models: AR, ARMA, ARIMA, Linear dynamic model (LDMs)
  • March 3, 2020 Jeongmin Lee : Recurrent Neural Networks: LSTMs, GRU
  Special Deep NN topics:
  • March 5, 2020. Convolutional Neural Networks (CNNs) (Tristan Maidment)
  Reinforcement learning:
  • March 24, 2020. Introduction to RL.
  • March 26, 2020. Markov decision processes, Dynamic programming.
  • ...

Course description
Lectures
TBA
Paper presentations
Projects

The goal of the field of machine learning is to build computer systems that learn from experience and that are capable to adapt to their environments. Learning techniques and methods developed by researchers in this field have been successfully applied to a variety of learning tasks in a broad range of areas, including, for example, text classification, gene discovery, financial forecasting, credit card fraud detection, collaborative filtering, design of adaptive web agents and others.

The objective of the Advances Machine Learning course is to expand on the material covered in the introductory Machine Learning course (CS2750). It focuses on special topics in ML such as exact and approximate inference in graphical models, dimensionality reduction and component analysis methods, latent variable models, models of documents and words, time series models, selected topic from deep neural networks and reinforcement learning. The course will consist of (student-lead) presentations and discussions. Students will be evaluated based on their participation in discussions, presentations and projects.

Prerequisites

CS 2750 Machine Learning, or the permission of the instructor.

In the first part of the course we will use readings from:

• Chris Bishop. Pattern Recognition and Machine Learning. Springer, 2006.

In addition we will use conference and journal paper readings that will be distributed electronically or in a hardcopy form.

Course webpage for CS2750, the introductory Machine Learning course from Spring 2019. It is the prerequisite of CS3750.

Readings will be assigned before the class at which the discussion on the topic takes place. Most of the readings will be electronic, however, some readings will be in the paper form or from the books. The students should always read the book chapter and paper assigned and be prepared for the class.

See the list of readings for different topics (currently incomplete)

Every student is expected to be in charge of two topics, present it, and lead the discussion on the topic during the class. The readings for each topic will be distributed electronically. The assignment of the papers will be discussed during the first two weeks of the course.
 

There are two projects for course. The first project (a group project) will be assigned and due in the middle of the semester. The final project (due at the end of the semester) and is more flexible: a student or a group (of a size at most 3) can choose from a set of topics/problems or propose his/her own topic to investigate. In general, the final project must have a distinctive and non-trivial learning or adaptive component.