Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/tomfran/ans-graph-compression
Graph compression using Asymmetric Numeral Systems
https://github.com/tomfran/ans-graph-compression
asymmetric-numeral-systems graph-compression webgraph
Last synced: about 2 months ago
JSON representation
Graph compression using Asymmetric Numeral Systems
- Host: GitHub
- URL: https://github.com/tomfran/ans-graph-compression
- Owner: tomfran
- License: lgpl-2.1
- Created: 2022-03-31T10:04:46.000Z (almost 3 years ago)
- Default Branch: main
- Last Pushed: 2024-07-05T19:00:54.000Z (6 months ago)
- Last Synced: 2024-07-06T00:32:55.105Z (6 months ago)
- Topics: asymmetric-numeral-systems, graph-compression, webgraph
- Language: Java
- Homepage:
- Size: 2.51 MB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# ANS Graph compression
This repository focuses on the application of asymmetric numeral systems to
large-scale graph compression.
Asymmetric numeral systems are a family of entropy encoders that obtains
compression quality comparable to arithmetic coding, thus optimal for a
given source of symbols, while having a decoding speed similar to Huffman
coding. They were presented by Jarek Duda and are heavily used in several
different scenarios, such as Linux kernels, Facebook Zstandard, JPEG XL
and many others.
This work presents an application of such encoders on large-scale web and
social graphs. The final proposed methodology is the result of three iterations
and combines asymmetric numeral systems, instantaneous codes and
patched frame of reference encoding.
The experimental results show how this methodology saves as much as 76
percent of space with respect to quasi-succinct representations and is
capable of storing graphs in as low as 3.5 bits per link.
This was my Master's thesis at the University of Milan, the complete work can be read [here](https://github.com/tomfran/ANS-Graph-compression/tree/main/thesis/thesis.pdf).