Bayesian-thinking / PersonalizedMultitaskLearning

Code for performing 3 multitask machine learning methods: deep neural networks, Multitask Multi-kernel Learning (MTMKL), and a hierarchical Bayesian model (HBLR).

PersonalizedMultitaskLearning

Code for performing 3 multitask machine learning methods: deep neural networks, Multitask Multi-kernel Learning (MTMKL), and a hierarchical Bayesian model (HBLR).

This code is still under construction!

Please note that this code is fresh from our research proejct. While all the functionality is here, we are still working cleaning up some aspects and providing proper documentation. We plan to complete this process before the publication of the associated paper. In the mean time, if you have any questions please email us at [e-mail] or [e-mail].

Please cite our work!

The journal paper associated with this code is currently under revision. For now, you can cite our earlier workshop paper:

Jaques N., Taylor S., Nosakhare E., Sano A., Picard R.,"Multi-task Learning for Predicting Health, Stress, and Happiness", NIPS Workshop on Machine Learning for Healthcare, December 2016, Barcelona, Spain. (*equal contribution) PDF BEST PAPER AWARD

Models in this code:

Multitask Neural Network

Multitask Multi-kernel Learning (MTMKL)

Hierarchical Bayesian Logistic Regression (HBLR)

Structure

Code structure

Wrappers are used to perform a grid search over hyperparameters

Input data format

.csv files

Assume csvs have columns for 'user_id', 'timestamp'

'Task dict list'

For the multi-task algorithms, we use a special data structure saved to a pickle file to represent the data from multiple tasks.

~~ Describe ~~

The code for generating files in this format give a .csv file is available in make_datasets.py