Smart Contracts Auto-generation for Supply Chain Contexts

Abstract

The introduction of blockchain technology into Supply Chain management has opened the possibility of faster and more secure transactions of commodities and services. As for every blockchain, Smart Contracts are the tool for controlling the transactions in blockchain-based supply chains. In this paper, we introduce a method for automating the implementation of natural language contracts into Smart Contracts in the Supply Chain context. The basic idea here is to extract information from a natural language contract using two Natural Language Processing (NLP) techniques, the Named Entity Recognition (NER) and Relation Extraction (RE), and then use this extracted information to automatically create a corresponding Smart Contract. This is an ongoing project, and we implemented the first phase of NLP, i.e., NER. The main issue we are facing here is the limited availability of annotated contract datasets. To tackle this challenge, we created an annotated legal contract dataset dedicated to the NER task. The dataset is analyzed with the deep learning method (BiLSTM) and transformer-based method (BERT). As per the generation of smart contracts, our approach consists of identifying meaningful entities and the relations between them and then representing them as business logic that can be directly incorporated into computer code as blockchain smart contracts.