Timezone: »

 
Poster
The Forget-me-not Process
Kieran Milan · Joel Veness · James Kirkpatrick · Michael Bowling · Anna Koop · Demis Hassabis

Wed Dec 07 09:00 AM -- 12:30 PM (PST) @ Area 5+6+7+8 #28
in Wednesday Posters »

We introduce the Forget-me-not Process, an efficient, non-parametric meta-algorithm for online probabilistic sequence prediction for piecewise stationary, repeating sources. Our method works by taking a Bayesian approach to partition a stream of data into postulated task-specific segments, while simultaneously building a model for each task. We provide regret guarantees with respect to piecewise stationary data sources under the logarithmic loss, and validate the method empirically across a range of sequence prediction and task identification problems.

Keywords: Information Theory · Online Learning · (Other) Probabilistic Models and Methods · Bayesian Nonparametrics · Multi-task and Transfer Learning · Time Series Analysis
[ PDF] 

Author Information

Kieran Milan (Google DeepMind)
Joel Veness (DeepMind)
James Kirkpatrick (Google DeepMind)
Michael Bowling (DeepMind / University of Alberta)
Anna Koop (University of Alberta)
Demis Hassabis (DeepMind)

More from the Same Authors