Paper accepted by ICML 2019 Sublinear Time Nearest Neighbor Search over Generalized Weighted Space. ICML 2019: 3773-3781, Yifan Lei, Qiang Huang, Mohan S. Kankanhalli, Anthony K. H. Tung

To build the project, use the following instructions:

mkdir build
cd build
cmake ..
make -j

Then run

./alsh --help

for help information.

A possible example to run could be:

./alsh -A fraction_recall_s2alsh -n 60000 -q 1000 -d 784 -D ../data/Mnist784/Mnist784.ds -Q ../data/Mnist784/Mnist784.q -W ../data/Mnist784/Mnist784_normal.w -G ../data/Mnist784/Mnist784_normal.gt -O output.out -U 3.140000 --data_hash_filename data_hash.dh --query_hash_filename query_hash.qh

The python scripts on scripts folder can be used to run the automatically run the precision_recall and fraction_recall experiments, e.g.

python ../scripts/run_ground_truth.py

python ../scripts/run_precision_recall.py

python ../scripts/run_fraction_recall.py

In order to run the scripts, the datasets/query should be put in the ./data folder, or the "./scripts/dataset_config.py" should be modified accordingly. Datasets link via onedrive

For more information, please contact [email protected]