Topositus will make it possible to combine cancer register data together while still maintaining privacy and protecting patient health data. This is the first proof of concept project.
Read about Opdan's research and innovation project
Data which includes various types of information are continuously being collected throughout the society and proper data management is necessary to ensure that no data leakage or theft occurs. This is particularly critical when managing personal data such as patient health data which can be at risk of being misused by nefarious parties.
Nikolai Opdan and his team realised that theoretical mathematical methods could be applied by industry and help address this risk and protect data from being leaked.
– We have implemented a new method called homomorphic encryption which allows for processing encrypted data. Information that is embedded within the data is hidden and unusable when encrypted. Access to this information requires one to process this data through decryption to access this information, which makes it impossible to steal, says Opdan.
Centralised data processing
Opdan further clarifies that their method will allow data to be processed while in its encrypted form, making the hidden information inaccessible and valueless if stolen. This could also allow centralised data processing in the future.
– This method allows data to be stored in its encrypted form at all times, even through the processing process. Data is a very valuable asset for a company and with this method, no one will be able to steal it even if they wanted to. By processing data that is encrypted, you could also process data from other sources and centralise a lot of data processing and allow for cloud-processed data in a more secure way, says Opdan.
According to Opdan, processing of a lot of data in Norway is costly as it must be done within the infrastructure of the country and prohibited to be sent abroad because of its sensitivity such as personal health information. Their method will allow for this centralising of processing and effectively reduce the cost and ability to use more than one data center thereby optimizing the use of the infrastructure.
First proof of concept project in cancer project
Opdan and his team are collaborating with Oslo University Hospital to research a cancer data set to gain more understanding on different cancers using machine learning. However, to be able to do this, more cancer register data is needed which is not possible to be easily acquired and combined today. Topositus will make it possible to combine cancer register data together while still maintaining privacy and protecting patient health data. This is the first proof of concept project.
– Combining different cancer registry data from other countries is a challenge due to the sensitivity of the data. People are also either unwilling or has no liberty to share this data. This process gives the possibility to combine and process encrypted data while maintaining the data in its safe, encrypted form. This would make the data pool large enough to do machine learning, says Opdan.
They are currently at the beginning of this first proof of concept project and if successful has the potential to solve the important problem pertaining to not just understanding this specific cancer but also potentially other diseases as well.
Can be used to detect fraud in the banking sector
Another field where their solution can be useful is the centralisation of data through this method can also be useful in fraud detection which is a very complex issue involving multiple stakeholders. To understand this issue and track the transactions, one would require information from all of them which is not feasible today.
– Each bank only has information about the transactions within that bank itself. The complexity of this issue is compounded with the involvement of several banks in different countries which cannot share transactional data with each other. Establishing a centralised data system would allow for the encrypted data from all the banks involved to be combined and the ability to see the whole picture and solving the fraud detection issue, says Opdan.
The bigger picture gained through the combination of the data in this centralised data system allows for the ability to look for patterns which can be advantageous in general data analysis.
– This allows for studying the data without knowing what to look for. We don’t look for small changes in the data but the structure in the datasets. So, it is an entirely different way of looking at the data. Every data analysis method has its advantages and disadvantages. Our algorithm in a way combines all of them and allows for doing data analysis without any assumptions, adds Opdan.
Read about how Opdan has worked with the UiO Growth House
Paving a new innovation journey
Opdan and his team started their innovation journey as a data analysis firm where they wanted to apply their original algorithm to the field of data science and find a place in industry that would have significant application potential. The challenge then was that access to this data is limited or completely absent. They restrategised and developed this new method to process encrypted data and sought out help from the UiO Growth House to set their idea in motion.
Due to the novelty of their method, Opdan and his team are getting advice from the UiO Growth House to find and prioritise the application of their technology which is the current challenge. They are relying on the business network and the networking platforms of the UiO Growth House to understand the specific challenges that companies are facing and learning about the potential specific application of their technology and how to communicate their ideas to these companies. The communication of this idea was an additional challenge due to the abstractness of their method.
– Trust is an essential factor. It is not common to tell other people about the problems they are facing in the company as their competitors can know their weak spots. Therefore, networking became extremely important to get on the inside of companies. Through networking platforms like the UiO Growth House Innovation Hangout, we got to meet a lot of people that were interested in our idea. We got a lot of help to pitch our complex idea as people need to understand our idea without risking them stealing it. It actually became a bigger challenge than implementing the software itself, says Opdan.
The networking through the UiO Growth House has also allowed them to be introduced and be part of the startup community in Norway as well as potential investors.
According to Opdan, the UiO Growth House Innovation Hangouts has been a great platform for him and his team to communicate with companies and building this company-university relationship. They have also gotten some essential seed funding for computers from the UiO Growth House to be able to develop this method which has been key to get Topositus going.
– To develop software, you need good computers. Because we have to have state-of-the-art computers in order to make state-of-the-art software. The dream is to become the first, biggest and leading firm by implementing our software where it is most useful. As data security becomes a more important aspect with the increase of the amount of data, the costs to manage this data with the techniques we have today is very significant. Our method/software would help reduce these costs and hopefully resolve a lot of issues that are out there, says Opdan.
They are also cooperating very closely with the Norwegian National Security Authority to ensure that the correct standards are put in place in a timely and orderly manner as they develop their technology. The high technical advancement level of their technology requires them to work across cross-professionally to ensure that Topositus reaches its maximum potential in the near future.
– We are working on different levels; we are working up to the governments to get new laws in place and with regulatory parties like the Norwegian National Security Authority to get them trust our algorithm and working with different industry partners to get the data. At the same time, we are developing our software to get it to fit to the needs of the industry, explains Opdan.