Back to blog
Blockchain & AI
Written by Shahnawaz Aziz
Harmony of Blockchain and AI
Data security, immutability and origin is of utmost importance for AI services provided via machine learning or deep learning. Data authenticity is key to an artificial neural network and various algorithms within it. Blockchain provides a trust-less environment for the parties involved without the knowledge of their identities. It allows participants to transfer assets across the internet without the need for any authentication. It’s a technology that enables peer-to-peer transactions with no requirement for a centralised authority.
At FarmaTrust we are exploring the relation between Blockchain and AI to solve challenges in the Pharma and Healthcare sector.
AI is a technology that gathers data and applies probability studies and behavioural patterns to create decision making processes for the machines. Implementing AI, computers are used for knowledge, reasoning, problem-solving, creating a perception, learning, planning and for the capability to manipulate or move objects.
AI is centralized and Blockchain is decentralized so where is the marriage between the two technologies?
Figure1. General Framework for supervised classification algorithms.
With AI, the outcome is probability based. Algorithmic processes are involved between the input and output which are generally not discussed between various stakeholders. With Blockchain, data is transparent, secure and irrefutable. However, Blockchain and AI can complement each other to provide better results.
AI relies on the data accuracy and authenticity to perform algorithmic computation for the result. Blockchain is well known for security and immutability of the data. Therefore, Blockchain will provide secure data which will support towards better AI models and results in more accuracy.
Artificial Neural Network (AAN) in AI uses a complex set of calculations and algorithms to decide the output for any given input. It is becoming increasingly difficult to track and process the micro-decisions taken by the hidden layers within the ANN powering the outputs. Also, there are concerns of misuse of the information as you can force the output by changing the input and thus influencing the training of the neural network. These problems can be solved by using the blockchain technology. Blockchain provides data audit trails,boosting the credibility of the organizations involved and increasing the trustworthiness. Blockchain can also help the companies to track back the route and trace the history of the data.
At present, AI runs on the centralized data repositories. Most of these repositories are in a public or a private cloud with the risk of losing data or getting corrupted. Blockchain can help in the data management by giving a low-cost solution. By using the blockchain technology, the data can be stored on millions of nodes. Multiples copies will make the data resilient. Moreover, the data will be time-stamped and chronologically maintained. Hence data recovery will be easy and more accurate.
The blockchain technology enables data sharing in a secure manner and thus allowing AI to share outcomes at large scale in a secure manner. Blockchain will enable the future data exchanges to securely craft and supply authentic data for more input processing in the ANN of the AI machines.
DEMAND FORECASTING USING NEURAL NETWORK FOR SUPPLY CHAIN MANAGEMENT
Ashvin Kochak1* and Suman Sharma1, Research Paper, ISSN 2278 – 0149 Vol. 4, No. 1, January 2015 © 2015 IJMERR. All Rights Reserved