Healthcare Cloud

Zero-knowledge computation for healthcare data

The Solution

Hallex is building a healthcare analytics platform for zero-knowledge computation. What this means is that the server or entity doing the healthcare analysis never actually sees the raw data. It is a system based on Fully Homomorphic Encryption that can load WebAssembly instructions (which most modern programming languages compile to) and run it on encrypted data. This allows a healthcare provider to perform arbitrarily complex analysis of medical information such as DNA or blood compositions without learning anything about the underlying data or even the result of the analysis. For the first time, analysis can be done without compromising the privacy of data owner.


Our aim is to unlock a marketplace previously inaccessible due to privacy concerns, regulation and access; now research institutions, large pharmaceutical companies or even start-ups can build machine learning models for healthcare since we have unlocked how anyone in the world can have their sensitive data such as DNA or medical history analysed without revealing any information to the company performing the analysis.


Hallex provides the platform that backs this marketplace, which includes the specialized server infrastructure that runs the machine learning models (running any algorithm under fully homomorphic encryption is a resource heavy computation). Everyone will be able to access the same providers on the same platform in a fully secure way as long as they have their own health records. With a platform like Hallex, we can finally unlock the potential of machine learning in healthcare, even on the most sensitive data.


The Problem We Are Solving And Why

Companies are currently rushing to build AI/Data Science models, but often find themselves blocked by access to data. Machine learning won't be useful for most healthcare applications before we have solved the data access problem.

  • At present, the companies that sequence and analyze DNA almost always take ownership of the data; compromising privacy of the most sensitive medical data
    • To solve this, many organizations resort to trying to anonymize data. However, this approach is fundamentally broken. Since a person's DNA string is unique to them, removing the person's name from the database won't allow them to hide the identity. The other popular approach is simply to give up ownership of your DNA and accept that companies offering genetic analysis will have access to it, which of course is not ideal.
  • The inability to preserve privacy of personal data also blocks collaboration across the healthcare and pharmaceutical industry, thus inhibiting progress and adoption
  • Currently DNA sequencing and analysis for private consumers is entirely tied together (companies that do genetic analysis provide everything). We disentangle these two steps by allowing the end-user to own their own personal data and choose from multiple providers to perform the analysis

How You Will Benefit

We believe we have the technology to improve healthcare, but the exchange between healthcare providers and patients is broken. Bioinformaticians will understand diseases better and consequently it will lead to more precise diagnostics and better treatments. However, this is not possible right now because they can't access the data.

We also believe every person will have their DNA sequenced in the future, but at the moment, no marketplace for analysis exists. There will be a significant need for interpreting the sequence DNA to unravel:

  • Instant risk score predicting likelihood of future diseases (forget the weeks/months wait time)
  • Personalised medicine - provide and receive better and more targeted healthcare
  • Specific food intolerances (example lactose) in order to suggest how you might improve your dietary choices for a better lifestyle
  • The importance of regular exercise (risk factor for stroke higher in some people than others)
  • Advancement in cures and medical treatments for deadly diseases (as a result of collaboration between big pharma/government organisations due to a fully secure data sharing)