Some of the latest software and datasets on which I participated
MUTE: A web-based peer-to-peer real-time collaborative editor
The software is distributed under the term of GNU Affero GPLv3 licence and is freely available at: https://github.com/coast-team/mute.
MUTE is a web-based text editing tool that allows users to edit documents collaboratively in real-time. It relies on the CRDT LogootSplit algorithm (https://inria.hal.science/hal-00903813/file/main.pdf) for the consistency maintenance. Existing collaborative editing editors rely on a centralised architecture, while MUTE relies on a peer-to-peer architecture. In this collaboration users maintain their data and share their data directly with their partners without transmitting their data through a central server. As a peer-to-peer system, our editor is highly scalable and resilient to faults and attacks. Our editor offers support for working offline while still being able to reconnect at a later time.
The influence of trust score on cooperative behavior in trust game
The software is freely available at: https://github.com/coast-team/trust_influence_analysis.
Implementation in zTree of four variations of the trust game: with and without partner identity and with and without explicit computation of user trust values based on the computational trust model proposed. Publication of the datasets collected during an experimental design with 30 people on the variations of the trust game and of the analyses of the collected data.
Trusternity
Implementation of Trusternity, an auditing scheme for public key transparency mechanisms using Ethereum blockchain. Trusternity is implemented as an extension to CONIKS (http://coniks.org), a key management system that provides transparency and privacy for end-user public keys. The extension to CONIKS is available at : https://github.com/coast-team/coniks-go. Definition of Ethereum contracts that publish the root of the signed Merkle tree (STR) of CONIKS server in Ethereum blockchain is available at: https://github.com/coast-team/trusternity-contract/
Prediction of quality classes of Wikipedia articles
Prediction of quality classes of Wikipedia articles using different machine learning techniques including feature engineering (e.g. random forest) and deep learning (e.g. RNN-LSTM). The software is freely available at: https://github.com/vinhqdang/wikipedia_analysis.
Measurement of real-time performance of collaborative editors
Evaluation of delays of several web-based real-time collaborative editing systems: GoogleDocs, Etherpad, Cryptpad, MUTE. The software is freely available at: https://github.com/vinhqdang/collaborative_editing_measurement.
