This is the official project page for "Accelerating Approximate Aggregation Queries with Expensive Predicates"
Please read the paper for full technical details.
Install the requirements with pip install ray scipy mystic numpy pandas tabulate tqdm matplotlib. Then run pip install -e inside the repo to install the abae python package.
Run bash make_directories.sh. Then run any corresponding .py file in to reproduce experiments in the paper. Note that we have precomputed all the oracle labels.