Abstract
We review some past work on the nascent but rapidly growing Algorithms with Predictions and discuss some work in progress in that area. Some of the work in progress we discuss is in using predictions with queues, including 1) queues with predictions in the setting of the power of two choices 2) queues with very small prediction-based advice. We also review the learned Bloom filter and explain how to make better use of the learned models by using multiple Bloom filters for different score ranges. Experiments show this partitioned learned Bloom filter can yield significantly better results.