Machine Learning and AI used at Coinbase
Machine learning is the science of getting computers to act without being explicitly programmed. In the past decade, machine learning has given us self-driving cars, practical speech recognition, effective web search, and a vastly improved understanding of the human genome. Machine learning is so pervasive today that you probably use it dozens of times a day without knowing it.
“Machine learning helps us balance risks for Coinbase, with flexibility for customers where we want them to have the best experience possible.”
— Soups Ranjan Director of Data Science Coinbase
This article covers yet another industry that depends on machine learning as the core dependency of their application. Coinbase is a secure platform that makes it easy to buy, sell, and store cryptocurrency like Bitcoin, Ethereum, and more. This US-based digital asset exchange Coinbase has revealed that it’s focused on building “state-of-the-art” machine learning (ML) technology with “efficient” execution for the crypto and blockchain-focused economy.
Let’s check out the exclusive application of Machine Learning at Coinbase. Coinbase has developed several high-performing ML models, including EasyML, Seq2Win, and EasyBlend. Companies like Coinbase that work in the field of cryptocurrency and blockchain are at very high risks relatively, thus they need robust and stable solutions to prevent fraud and illegal activities which may be practiced by hackers and cybercriminals.
EasyML at Coinbase
Easy Machine Learning presents a general-purpose data flow-based system for easing the process of applying machine learning algorithms to real-world tasks. Machine learning algorithms have become the key components in many big data applications. The key barriers come from not only the implementation of the algorithms themselves but also the processing for applying them to real applications which often involve multiple steps and different algorithms. Also, EasyML is a no-code platform to build Powerful Machine Learning Models.
Tabular data at Coinbase consists of features like age, location, crypto-reserve, etc. These features are fed to EasyML. EasyML auto-detect features and returns various interpretations and visualization.
Seq2Win at Coinbase
Apart from tabular data, companies have to also keep track of sequential data. Sequential data at cryptocurrency platforms may consist of queues of events with timestamps, transaction logs, deploying airdrops, and much more. Sequential data is the data in which the order of the events or the records matters. Seq2Win application is used to process such sequential records, build models and provide insights to the company.
EasyBlend
EasyBlend was created to allow the developers to combine models from EasyML and models from Seq2Win efficiently. Easy blend trains and combines these models under the hood using Linear Blending
AWS Sagemaker
Coinbase uses machine learning models on Amazon SageMaker to help with fraud prevention, identity verification, and large-scale compliance. Using Amazon SageMaker reduced model training times from 20 hours to 10 minutes. Services like Sagemaker makes the training cost less than 90 percent and inference costs less than 70 percent.
Generative models
Generative models are a type of deep learning method specialized in generating synthetic data that matches the distribution of a training dataset. In our scenario, imagine that we train a generative model in the distribution of the link order book in Coinbase in order to generate new orders that match the distribution of the real order book.
These are some of the various Applications of Machine Learning in the world of Cryptocurrency. Machine Learning as Data Engineering has exponentially evolved over the last decade. Machine Learning is capable of finding patterns in the data or Big Data which is not feasible to interpret for statisticians too.
Thank you!
#arthbylw #vimaldaga #righteducation
#educationredefine #rightmentor
#worldrecordholder #ARTH #linuxworld
#makingindiafutureready #righeudcation