pure-ldp 1.0.2

Simple pure LDP frequency oracle implementations

Pure-LDP

pure-LDP is a Python package that provides simple implementations of pure LDP frequency oracles detailed in the paper "Locally Differentially Private Protocols for Frequency Estimation" by Wang et al

The package has implementations of all three main techniques:

1. (Optimal) Unary Encoding - Under pure_ldp.unary_encoding.ue_client and pure_ldp.unary_encoding.ue_server
2. (Summation/Thresholding) Histogram encoding - Under pure_ldp.histogram_encoding.he_client and pure_ldp.histogram_encoding.he_sever
3. (Optimal) Local Hashing - Under pure_ldp.local_hashing.lh_client and pure_ldp.local_hashing.lh_server

The package also includes an implementation of the heavy hitter algorithm Prefix Extending Method (PEM)

• This is under pure_ldp.prefix_extending.pem_client and pure_ldp.prefix_extending.pem_server

There is also support for the frequency oracle Hadamard Response, the code implemented for this is simply a pure-LDP wrapper of hadamard_response

• This is under pure_ldp.hadamard_response.hr_client and pure_ldp.hadamard_response.hr_server

Installation

Use the package manager pip to install.

pip install pure-ldp

To upgrade to the latest version

pip install pure-ldp --upgrade

Requires numpy, scipy, xxhash, bitarray and bitstring

Usage

import numpy as np
from pure-ldp.local_hashing.lh_client import LHClient
from pure-ldp.local_hashing.lh_server import LHServer

# Using Optimal Local Hashing (OLH)

epsilon = 3 # Privacy budget of 3
d = 4 # For simplicity, we use a dataset with 4 possible data items

client_olh = LHClient(epsilon=epsilon, d=d, use_olh=True)
server_olh = LHServer(epsilon=epsilon, d=d, use_olh=True)

# Test dataset, every user has a number between 1-4, 10,000 users total
data = np.concatenate(([1]*4000, [2]*3000, [3]*2000, [4]*1000))

for item in data:
    # Simulate client-side privatisation
    priv_data = client_olh.privatise(item)

    # Simulate server-side aggregation
    server_olh.aggregate(priv_data)

# Simulate server-side estimation
print(server_olh.estimate(1)) # Should be approximately 4000 +- 200

See example.py for more examples.

Acknowledgements

1. Some OLH code is based on the implementation by Tianhao Wang: repo
2. The Hadamard Response code is just a wrapper of the k2khadamard.py code in the repo hadamard_response by Ziteng Sun

Contributing

If you feel like this package could be improved in any way, open an issue or make a pull request!

License

MIT