PR/CV Reading group

Please note that the schedule for the first cycle is very tentative, since there was little consensus over the type of papers that people wanted to discuss. Let's take a part of the first meeting to discuss what everyone would like to get out of these meetings.

You can find the Google Calendar for these meetings and the CoffeeTalks here

Cycle 1 (Spring 2014): Deep Learning

Recently learned deep architectures, such as convolutional neural networks, deep boltzman machines and deep belief networks, have shown large performance improvements over the state of the art on many prediction tasks. In this cycle we will investigate how these methods work and how and when to apply them in practice.

A related CVPR tutorial (link by Wenjie).

Cycle 2 (Fall 2014): Gaussian Processes

Related slides (link by Andre).

Cycle 3 (Spring 2015): Graphical Models

Cycle 4 (Summer 2015): Graphical Models, Exponential Families and Variational Inference

During the summer (2015) we will be reading the monograph "Graphical Models, Exponential Families and Variational Inference" by Wainwright and Jordan.

Cycle 5 (Fall 2015): Submodularity

Cycle 6 (Spring 2016): Learning bounds for regression and classification

List of suggested future topics

• Markov Chain Monte Carlo methods
• Gaussian Processes
• Learning bounds for regression and classification
• Manifolds and Riemannian Geometry
• Machine Learning with Privacy Constraints
• Big Data (as suggested by Marco)

Colloquium Slides

Thijs van Ommen - Inconsistency of Bayesian inference when the model is wrong, and how to repair it (slides)