Lifelong Open-Ended Probability Predictors

TMLR Paper6430 Authors

Abstract:

We advance probabilistic multiclass prediction on lifelong streams of
items. A (learner) predictor must provide item probabilities, adapting
to significant non-stationarity, including new item appearances and
frequency changes. The predictor is not given the set of items that it
needs to predict before hand, and moreover the set can grow unbounded:
the space-limited predictor need only track the currently salient
items and their probabilities.
We develop Sparse Moving Average techniques (SMAs), including
adaptations of sparse EMA as well as novel queue-based methods with
dynamic per-item histories. For performance evaluation, to handle new
items, we develop a bounded version of log-loss. Our findings, on a
range of synthetic and real data streams, show that dynamic
predictand-specific (per connection) parameters, such as learning
rates, enhance both adaptation speed and stability. Code is provided.

Submission Type: Long submission (more than 12 pages of main content)

Assigned Action Editor: ~Martin_Mundt1

Submission Number: 6430