Building trust through technology: Blockchain and Machine Learning
Blockchain is more than just a technology trend powering the parallel universe of bitcoins. Expert Jerome from ING Bank explains what it can do in combination with another IT trend: machine learning. The result is greater security, increased trust and broader application possibilities.
Just recently, the price of a bitcoin went through the roof – again. The soaring price could be due to the current uncertainty in the world and investors gravitating toward safe investments. Blockchain, the technology behind cryptocurrency, provides the needed security and trust. But digital money is just one of the many applications at which this technology excels. Expert Jerome explores, among other things, what benefits the trust delivered by blockchain technology could have for an online bank. He works as a business intelligence specialist and DevOps engineer for ING Germany – one of the many exciting IT jobs at this digital challenger bank. And trust, as we all know, is the most important asset for a bank.
A question of trust
Trust is at the top of the agenda right now. Not just for banks, in which customers entrust their funds and financial information, but also for society in general. On the one hand, there are the breathtaking possibilities of digital technology – such as machine learning and artificial intelligence. On the other hand, there is the fear that many people have of being overwhelmed by it. Compromised data privacy, cybercrime and digital tracking all feed into people’s mistrust of innovation and are leading to the tightening of data protection policies. This is precisely where blockchain can play a crucial role in ensuring that progress does not fail because of data protection concerns. “The central idea of blockchain is this: How can we manage to bring people together on a technological basis? Because today, after all, people are drifting further and further apart,” Jerome explains. More concretely, “You have a huge problem today if you want to use and process large amounts of data. A simple example of this is medicine. We have an infinite amount of data there, but nobody can do anything with it because it's simply not allowed. And then comes the consideration: Can we use blockchain here to establish data security, yet still process the data?” The distributed ledger – a core concept behind blockchain – creates a reliable protocol that makes the execution and documentation of processes and transactions objective, anonymous and transparent for all parties involved. Manipulation is impossible, because all previous transactions are archived in the blockchain, mathematically validated, and cryptographically protected.
Safeguarded identities, secure learning
In the context of a bank, one obvious use case is the safeguarding of identities. Jerome: “It's all about wanting to track who is doing something with my data and where this data is coming from.” If blockchain could one day be used to ensure that a certain action was actually carried out by customer XY, it would make quite a few processes – from opening accounts to purchasing financial products – more efficient. Before that, he believes internal use cases will roll out first. After all, the identities of employees could also be validated in this way, e.g. which is attractive for IT systems and access rights. “Until now, a lot of this has been done via certificates and cumbersome processes. If there were a digital identity architecture underneath, then you could move much more freely between systems,” says Jerome. It gets even more exciting when the basic technology of blockchain is brought together with machine learning. This fairly new idea is one of his specialties. He is a business information scientist and devoted his university thesis to the topic. Machine learning can be used to gain highly practical insights from data, such as evaluations of patient data or analyses of customer behavior. Using script-based smart contract logic, blockchain now offers the ability to design data access in a way that keeps user identities anonymous – and validated at the same time. Add in the concept of federated learning, and learning can even take place in a decentralized way on users' devices, be they PCs, smartphones or smart devices like fitness trackers. Local data is used to create partial models, which are then anonymized centrally and merged into an overall model. “The data doesn't leave the cell phone at all, for example,” Jerome says, explaining the approach. In addition, thanks to the increased variety of data, distributed learning prevents the statistical error of overfitting when training the models.
Technology and more
What are the technologies that Jerome considers particularly suited to this? It is important that only “closed” blockchains come into question. Unlike the “open” bitcoin blockchain, no proof-of-work is required to generate blocks. This also eliminates the considerable (and often criticized) amount of energy needed, e.g. when mining bitcoins. The closed blockchain could be a consortium blockchain, in which a limited circle of participating institutions or companies could accredit new participants, without allowing anyone to unilaterally change the modalities of consensus building and logging. For the “identity management” use case, he finds Hyperledger Indy particularly interesting. This open-source framework is used to manage identities and relationships between actors in a decentralized manner, across platforms and applications. “Ultimately, the challenge is not so much the technology, but bringing different people to the table,” Jerome adds, which is the next challenge he aims to tackle.
For developers, who want to further develop: IT jobs at ING
Jerome will soon be devoting himself even more intensively to his specialty, as he is moving to an IT project at ING that deals with digital identities: “Blockchain – it’s just getting started at ING!” It is one of many exciting topics that IT professionals can sink their teeth into at ING, today and in the future – with plenty of freedom for personal interests. There are also other advantages for ambitious IT professionals: the agile way of working, a generous training budget, a global IT community, and the job security of a global corporation. ING’s offices in Frankfurt am Main and Nuremberg are constantly seeking new employees. If you are interested in helping to shape the digital future of banking, you can find out everything you need to know about IT opportunities at ING on our careers page.