Decentralized ML: Part I (Introduction)


The first part of the decentralized machine learning series focuses on understanding the need for decentralization using an intuitive example.

Photo by Clint Adair on Unsplash

This is the first part of the series of blogs on the topic of Decentralized Machine Learning. Throughout the series, I will introduce the field of decentralized machine learning, explain the need for decentralization, cover some tools and techniques used to achieve decentralization, and end the series by providing a detailed account of the current research status and open problems in the field. I haven’t currently decided on the number of parts I plan to provide in the series but stay tuned for more information.

Join my Quora space, for regular news and updates about decentralized machine learning. Being extremely experienced in building and researching the decentralized machine learning systems, I will be posting original content about the latest progress in academia and the industry. Visit my personal website for more details about my education, skills, experiences, and credentials.

Photo by from Pexels

Machine learning and data science as a whole is an extremely booming field at present as it’s capable of providing powerful solutions that might be impossible to think of using the rule-based programs. However, data is the greatest bottleneck of any machine learning system. Today, although we might be able to think of really innovative machine-learning-based solutions, it won’t be practically possible to implement any of them if the relevant data isn’t available.

Photo by Pixabay from Pexels

Although there are a lot of ways to acquire publically-available data from the web, it’s practically impossible to acquire the data which might be private to the users or organizations. And according to me, our capability of indirectly utilizing the private data while upholding the principals of security and privacy will actually decide the scope of machine learning in the future. Although it’s practically possible to build a ton of solutions using the publically available data, we’ll hit the dead-end at some point.

Photo by Charles Deluvio on Unsplash

More importantly, the unavailability of private data imposes various limitations on the type of solutions we can provide. For example, the purchase recommendation systems that we frequently notice on the e-commerce websites provide solid suggestions and make our lives easier. But as it’s only trained on the purchase practices of other users and the data that is only private to the organization, there are a few limitations that we barely notice. Think about a scenario in which a lot of people purchased a 20 USD laptop sleeve and a 2000 USD laptop at the same time. So if you were to only purchase a 20 USD laptop sleeve in your cart, there’s a high chance that the system might recommend you to buy a 2000 USD laptop. This might get annoying at times, but it cannot be resolved unless the e-commerce organization can get insights about the budget or bank balances of the buyer. If the e-commerce recommender knew that you don’t have enough balance in your bank account or enough limit on your credit card, it wouldn’t have made a recommendation to purchase a 2000 USD laptop. In a similar manner, even banks can provide more customized products, services, and offers for the customers if they can get an insight into their shopping preferences from the e-commerce websites. However, to ensure the privacy of their users, the organizations cannot directly share the data with each other.

Photo by Terry on Unsplash

And hence, although machine learning is a booming field at present, it might not be necessarily relevant in the future. Although we have made tremendous progress in the fields of machine learning and deep learning, there are a lot of problems that need to be solved, with the decentralization of data being the immediate one. Although a lot of research has been done in the field in recent years there are a lot of problems (security, scalability, architecture, etc.) that still remain unsolved.

In the few upcoming posts, I plan to introduce the tools and techniques that can be used to achieve decentralization in a machine learning setting. In addition to the introduction, I will be providing blogs and resources toward the end of the article which will help you to expand on the information provided. Look out for Decentralized ML: Part II (Blockchain Basics). Thanks for reading!

Again, if you wish to get regular news and updates about decentralized machine learning, please follow my space on Quora. Based on my experience as a researcher, I will be sharing original content relevant to industrial and academic progress in the field of decentralized machine learning.

Get Best Software Deals Directly In Your Inbox

Decentralized ML: Part I (Introduction) was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.