This is the official repository for "Best Arm Identification with Fixed Budget: A Large Deviation Perspective", NeurIPS 2023
StatsBase, JLD2.jl, IterTools, StatsPlots.jl, CSV, DataFrames
Setting | Description |
---|---|
1 | one group of suboptimal arms |
2 | two group of suboptimal arms |
3 | linear |
4 | concave |
5 | convex |
6 | stair |
We fix K∈{10,20,40}
for setting_id∈{1,...,5}
and K∈{15,21,55}
for setting_id=6
, while varying the budget T
.
Each experiment is repeated N=40,000
times and the averaged error probability δ
is reported.
See the comments in experiments.jl and viz.jl for more details.
- Run the experiment:
julia -O3 -p8 experiment.jl {setting_id} {M}
, where-p8
means to use 8 processors - Plot the results:
julia -O3 viz.jl {setting_id} {M}
Name | Abbrev. | Description |
---|---|---|
Continuous Rejects with aggressive rate | CR-A | |
Continuous Rejects with conservative rate | CR-C | |
Successive Rejects | SR | (Audibert, et al., COLT'10) |
Sequential Halving | SH | (Karnin, et al., ICML'13) |
Uniform Sampling | RR |