Course on Computational Interaction with Bayesian Methods

Nikola Banovic, Per Ola Kristensson, Antti Oulasvirta, John Williamson

0900 - 1720 Wednesday 8 May 2019, Glasgow, UK

Launch the notebooks on Binder

No need to install anything: thanks to the amazing Binder service, you can open and run these notebooks directly on the web. Just click the link above to launch a VM.

Notebooks

0900-1020 01_intro_to_bayesian_methods/ Introduction to Bayesian methods in HCI and Bayesian filtering to estimate state
1100-1220 02_decoding_symbols/
1400-1520 03_bayesian_optimisation/
1600-1720 04_modeling_behavior/

Topic

The course focuses on optimization and inference and on applying these techniques to concrete HCI problems. The course will specifically look at Bayesian methods for solving decoding, adaptation, learning and optimization problems in HCI. The lectures center on hands-on Python programming interleaved with theory and practical examples grounded in problems of wide interest in human-computer interaction.

Instructors

The following faculty members will teach the course:

Nikola Banovic, University of Michigan, USA
Per Ola Kristensson, University of Cambridge, UK
Antti Oulasvirta, Aalto University, Finland
John Williamson, University of Glasgow, UK

Local install instructions

If you are not using mybinder.org, then you can download and install a local version:

Install Anaconda 3.6 for your platform if you don't already have it installed (note Python 3.7 currently has a conflict with gpyopt)

Clone the repository somewhere on your machine

  git clone https://github.com/johnhw/chi_course_2019.git

At the terminal, enter the directory where you cloned the repo
Create a new conda environment with the prerequisites
```
  conda env create -f environment.yml
```
Activate the environment with
```
  conda activate chi-course-2019
```
Start the notebook server with
```
jupyter notebook
```
and then open index.ipynb